(Received for publication, March 7, 1997)
From the Department of Pathology, Washington University School of Medicine, St. Louis, Missouri 63110
Glycoproteins, such as the glycoprotein
hormone lutropin (LH), bear oligosaccharides terminating with the
sequence SO4-4GalNAc1,4GlcNAc
1,2Man
(S4GGnM)
and are rapidly removed from the circulation by a receptor present in hepatic endothelial cells and Kupffer cells. Rapid removal
from the circulation is essential for attaining maximal hormone
activity in vivo. We have isolated a protein from rat liver
which has the properties expected for the S4GGnM-specific receptor
(S4GGnM-R). The S4GGnM-R is closely related to the macrophage mannose
receptor (Man-R) both antigenically and structurally. At least 12 peptides prepared from the S4GGnM-R have amino acid sequences that are
identical to those of the Man-R. Nonetheless, the ligand binding
properties of the S4GGnM-R and the Man-R differ in a number of
respects. The S4GGnM-R binds to immobilized LH but not to immobilized
mannose, whereas the Man-R binds to immobilized mannose but not to
immobilized LH. When analyzed using a binding assay that precipitates
receptor ligand complexes with polyethylene glycol, the S4GGnM-R is
able to bind S4GGnM-bovine serum albumin (S4GGnM-BSA) conjugates
whereas the Man-R is not. In contrast both the S4GGnM-R and the Man-R
are able to bind Man-BSA. Monosaccharides that inhibit binding of
Man-BSA by the Man-R enhance binding by the S4GGnM-R. Oligosaccharides
terminating with S4GGnM and those terminating with Man are bound at
independent sites on the S4GGnM-R. The S4GGnM-R present in hepatic
endothelial cells may account for clearance of glycoproteins bearing
oligosaccharides terminating with S4GGnM and glycoproteins bearing
oligosaccharides terminating with either mannose, fucose, or
N-acetylglucosamine.
Asn-linked oligosaccharides present on the glycoprotein hormones
lutropin (LH)1 and thyrotropin (TSH)
terminate with the sequence
SO4-4GalNAc1,4GlcNAc
1,2Man
(S4GGnM), whereas those
on follitropin and chorionic gonadotropin (CG) terminate with the
sequence Sia
2,3/6Gal
1,4GlcNAc
1, 2Man
(1-5). We have
proposed that the sulfated oligosaccharides present on LH and TSH are
critical for the expression of full biologic function by these hormones
(6-9). Terminal GalNAc-4-SO4 does not influence binding to
or activation of the LH/CG receptor itself (10) but does have a marked
impact on the circulatory half-life of LH following secretion (7, 11)
due to recognition of the sulfated oligosaccharides by a receptor
expressed at the surface of hepatic endothelial cells and Kupffer cells
(11, 12). The rapid removal of LH from the circulation in conjunction
with its release from granules in response to gonadotropin releasing
hormone accounts for the episodic rise and fall in hormone levels seen in the circulation. Since the LH/CG receptor is a G-protein-coupled receptor, which rapidly becomes refractory to further stimulation following ligand binding (13-15), episodic stimulation may provide for
maximal activation during the preovulatory surge in circulating LH
levels. TSH shows similar properties with respect to half-life and
receptor activation (16-19).
Glycoproteins bound by the S4GGnM-specific receptor are subsequently
transported to lysosomes and degraded. There are roughly 600,000 S4GGnM-specific binding sites at the cell surface of hepatic endothelial cells, which bind LH through its sulfated oligosaccharides with an apparent Kd of 2.7 × 107
M. Binding is pH-dependent, requiring a pH > 5.0, but does not require Ca2+ (12). The location of the
sulfate in the 4-position is critical since glycoconjugates bearing
oligosaccharides terminating with the sequence
SO4-3GalNAc
1,4GlcNAc
1,2Man
(S3GGnM) are not bound by hepatic endothelial cells. We have now identified and isolated a
glycoprotein from rat liver that has the properties expected for the
receptor, which mediates removal of LH from the circulation on the
basis of its sulfated oligosaccharides.
Analytical Procedures
Protein concentrations were determined using the Bio-Rad protein
assay kit (Bio-Rad) or ISS Protein-Gold (Integrated Separation Systems). Polyacrylamide gel electrophoresis in the presence of sodium
dodecyl sulfate (SDS-PAGE) was performed according to Laemmli (20).
Following separation by SDS-PAGE on 5% or 7.5% acrylamide gels,
proteins were transferred electrophoretically to polyvinylidene difluoride membranes using CAPS buffer as described by Matsudaira (21)
for amino-terminal sequence determination, peptide mapping, and
detection with specific antisera. Proteins detected by antisera were
developed using 125I-F(ab)2 goat anti-rabbit
IgG.
Affinity Columns
Wheat Germ Agglutinin (WGA)-SepharoseWGA (Sigma) was dissolved in 100 mM NaHCO3, 100 mM GlcNAc, 200 mM NaCl, pH 8.4, at a concentration of 2.0 mg/ml and added to cyanogen bromide-activated Sepharose 4B (Sigma) at a ratio of 5.0 mg/ml Sepharose 4B. Following 2 h of rotation at room temperature, >95% of the WGA was coupled. Remaining active sites were quenched by incubation overnight at 4 °C in 200 mM glycine, pH 8.0. The WGA-Sepharose was washed successively with 0.1 M sodium acetate, 1.0 M NaCl, pH 4.0, and 0.1 M borate, 1.0 M NaCl, pH 8.0. WGA-Sepharose was stored in 20 mM Tris-HCl, 0.15 M NaCl, 2.0 mM CaCl2, pH 7.8, containing 0.2% NaN3.
bLH-SepharosebLH-Sepharose was prepared in the same fashion with the following modifications. bLH was dissolved in 100 mM NaHCO3, 200 mM NaCl, pH 8.4, at a concentration of 3.8 mg/ml and added to cyanogen bromide-activated Sepharose 4B at a ratio of 6.5 mg/ml Sepharose. The coupling efficiency was >95% after 2 h at room temperature. Remaining active sites were quenched by incubation at 4 °C overnight in 1.0 M ethanolamine, pH 8.3. bLH-Sepharose was stored in 20 mM Tris HCl, 0.15 M NaCl, 2.0 mM CaCl2, pH 7.8, containing 0.2% NaN3.
Radiolabeling
SO4-4GalNAc1,4GlcNAc
1,2Man
-(CH2)8COO-bovine
serum albumin (S4GGnM-BSA),
SO4-3GalNAc
1,4GlcNAc
1,2Man
-(CH2)8COO-bovine serum albumin (S3GGnM-BSA), Mannose-bovine serum albumin (Man-BSA), or
purified S4GGnM receptor, 5-10 µg, were dissolved in 50 µl of 20 mM Tris-HCl, 0.15 M NaCl, pH 7.4, and incubated
for 15 min on ice with a single IODOBEAD (Pierce) and 2.5 µCi of
125I (Amersham, IMS30). Labeled proteins were separated
from reaction products by gel filtration on Sephadex G-10 (Pharmacia
Biotech Inc.) in 20 mM Tris-HCl, 0.15 M NaCl,
pH 7.8, containing 1 mg/ml bovine serum albumin. The fractions
containing 125I-labeled product were pooled and stored at
20 °C for no longer than 2 months. F(ab
) goat anti-rabbit IgG
(250 µg) was iodinated in the same fashion using 500 µCi of
125I in 500 µl of 20 mM NaPO4,
0.15 M NaCl, pH 7.5, and incubated 1 h at room
temp.
Binding Assays
Binding assays (total volume of 150 µl) contained 3-5 ng of
S4GGnM receptor 125I-S4GGnM-BSA (2-3 × 105 dpm),
and 90 µg of hyaluronic acid and/or 90 µg of fucoidin in 20 mM Tris-HCl, 0.15 M NaCl, 2 mM
CaCl2, 1% (w/v) Triton X-100, pH 7.8. Hyaluronic acid is a
weak inhibitor of S4GGnM-BSA binding, whereas fucoidin is a potent
inhibitor (12). Incubations were performed in a 10 × 75-mm glass
tube at room temperature for 30 min. The reactions were terminated by
adding 1.5 ml of ice-cold 10% (w/v) PEG 8000 (Sigma) in 20 mM Tris-HCl, 0.15 M NaCl, 2 mM CaCl2 and mixing. After 30 min on ice, precipitated
125I-S4GGnM-BSA·S4GGnM receptor complexes were collected
by vacuum filtration on Whatman GF/C filter discs, which had been
soaked in 20 mM Tris-HCl, 0.15 M NaCl, 2 mM CaC12, 5 mg/ml bovine serum albumin. The
filters were washed twice with 1.5 ml of ice-cold 20 mM
Tris-HCl, 0.15 M NaCl, 2 mM CaCl2,
10% (w/v) PEG 8000, and the amount of 125I determined by
counting the filter in a -counter. In the absence of added S4GGnM
receptor, <5% of the added 125I-S4GGnM-BSA was captured
on the filter. One unit of activity is defined as the amount of S4GGnM
receptor that is able to precipitate 1 ng of S4GGnM-BSA in the presence
of hyaluronic acid above that precipitated in the presence of both
hyaluronic acid and fucoidin.
S4GGnM Receptor Isolation
Step 1: HomogenateHarlan Sprague Dawley rats, 150-200 g each, were anesthetized and heparinized and their livers perfused with ice-cold 20 mM PO4, 0.15 M NaCl, pH 7.5, through the portal vein. Each liver was suspended in 20 ml of 0.25 mM EDTA, 0.02% NaN3 (w/v) brought to pH 7.8 with solid NaHCO3 and containing 50 units/ml aprotinin. The suspension was homogenized with three 20-s bursts of a Polytron homogenizer (Brinkman) at a setting of 5. Alternatively, 50 frozen rat livers weighing 300 g (Pel-Freez) were ground while frozen using a meat grinder and suspended in 400 ml of 0.25 mM EDTA, 0.02% NaN3 (w/v) brought to pH 7.8 with solid NaHCO3 and containing 50 units/ml aprotinin. The suspension was homogenized with four 15-s bursts of a Polytron homogenizer (Brinkman) at a setting of 8.
Step 2: Triton X-100 SolubilizationSufficient 25% (w/v) Triton X-100 (Boehringer Mannheim) was added to bring the concentration of the homogenate to 10%, and the pH was adjusted to 7.7 with 1.0 M Tris-HCl, pH 7.8. After stirring for 1 h at 4 °C, the Triton extract was passed through cheesecloth to remove connective tissue and sedimented at 7100 × g for 20 min.
Step 3: PEG PelletSolid PEG 8000 (Sigma) was added to the supernatant to a final concentration of 10% (w/v). The extract was stirred for 15 min at 4 °C and then allowed to stand for 30 min. Precipitated proteins were collected by sedimentation at 7100 × g for 90 min. The supernatant was discarded.
Step 4: Triton X-100 Solubilization of PEG PrecipitateThe PEG precipitate was resuspended by vigorous stirring for 30 min in 20 mM Tris-HCl, 0.2 M NaCl, 0.05 M NaN3, 3 mM CaCl2, 1% (w/v) Triton X-100, pH 7.8, 20 ml/liver.
Step 5: WGA-SepharoseSolubilized proteins from the PEG precipitate were incubated with WGA-Sepharose (1-2 ml of WGA-Sepharose/liver) overnight at 4 °C. Unbound proteins were removed by washing on a sintered glass funnel with 20 mM Tris-HCl, pH 7.8, 150 mM NaCl2, 1% Triton X-100. Bound glycoproteins were subsequently eluted using the same buffer containing 300 mM GlcNAc.
Step 6: bLH-SepharoseThe WGA-Sepharose eluate was incubated with 0.5 ml of bLH-Sepharose/liver overnight at 4 °C with rotation. Unbound material was removed by washing on a sintered glass funnel with 20 mM Tris-HCl, pH 7.8, 150 mM NaCl, 2 mM CaCl2, 1% Triton X-100. The bLH-Sepharose was then eluted with 50 mM sodium acetate, pH 4.0, 0.1% Triton X-100. The eluate was immediately adjusted to pH 7.0 by addition of 1.0 M Tris base and the volume reduced using a Centriprep-10 (Amicon).
Step 7: Mannan-Sepharose or Mannose-SepharoseThe bLH-Sepharose eluate was incubated with either mannan-Sepharose (0.05 ml/liver) or mannose-Sepharose (0.05 ml/liver) and the unbound fraction taken. Bound proteins were eluted from the mannan-Sepharose and mannose-Sepharose by successive incubation with 50 mM galactose in 20 mM Tris, 150 mM NaCl, 0.05% Triton X-100, and 0.2 mM Pefabloc SC (Boehringer Mannheim) adjusted to pH 7.5 and 200 mM mannose and 5 mM EDTA in the same buffer.
Preparation of Rabbit Antisera to the S4GGnM Receptor
Polyclonal antisera to the S4GGnM receptor were raised in New Zealand White rabbits by immunization with the 180-kDa protein band isolated from an SDS-polyacrylamide gel. The protein band was excised, and, after extracting the gel pieces with 95% ethanol to reduce the SDS content, the gel was lyophilized. The dried gel was pulverized using a mortar/pestle and then emulsified with saline by passing through a series of successively smaller needles ranging from 18 to 23 gauge. The emulsified gel containing 5-10 µg of protein was added to TDM-emulsion (RIBI) and injected intramuscularly into the hind leg and subcutaneously at four separate sites. The rabbit was boosted in the same manner 3 weeks later and sera obtained 8-10 days later. The rabbit was subsequently boosted with antigen as required to maintain the titer of the antisera.
Radioimmunoassay for S4GGnM Receptor
The affinity-purified S4GGnM receptor was labeled with
125I as described above. An antibody saturation curve was
established using a constant amount of radiolabeled receptor (5-10 ng,
3 × 105 cpm) and increasing amounts of antisera.
Following an overnight incubation at 4 °C, protein A-Sepharose
antigen-antibody complexes were washed twice with 20 mM
phosphate-buffered saline, pH 7.5, containing 0.1% BSA (w/v) and
counted in a counter. Standard inhibition curves were constructed
using a constant amount of radiolabeled receptor, sufficient antisera
to precipitate 50-70% of the 125I-S4GGnM-R added, and
increasing amounts of unlabeled receptor that had been quantitated by
amino acid analysis. The amount of receptor in "unknown" samples
was then determined by comparison to the standard inhibition curve.
We previously identified a receptor in rat liver that can account for the rapid removal of native LH bearing Asn-linked oligosaccharides terminating with the sequence S4GGnM from the circulation (12). The S4GGnM-R is located predominantly in hepatic endothelial cells and Kupffer cells and displays a high degree of specificity, recognizing S4GGnM-BSA but not S3GGnM-BSA. Fucoidin, a sulfated polysaccharide, inhibits binding of S4GGnM-BSA and LH by the receptor, whereas other sulfated and anionic polysaccharides do not inhibit binding or require much higher concentrations to do so. Glycoproteins bound to the S4GGnM-R are internalized, transported to lysosomes, and degraded. Binding is pH-dependent, requiring a pH above 5-6, and is not dependent on divalent cations even though divalent cations do enhance binding.
We established conditions that allowed us to detect and solubilize a binding activity with the properties we had described for the S4GGnM-R. Parenchymal cells (hepatocytes) and endothelial/Kupffer cells, prepared by collagenase perfusion as described previously (12, 22), were disrupted by Dounce homogenization, and the nuclei and unbroken cells were collected by centrifugation. Soluble and total membrane fractions were obtained by sedimentation onto a 65% sucrose cushion. Fractions were brought to a final concentration of 1% (w/v) Triton X-100 and assayed for S4GGnM-BSA binding using the PEG precipitation assay. Binding activity was confined to the membrane fraction and the pellet containing nuclei and unbroken cells. None was found in the soluble fraction. As much as 80% of the binding activity was recovered in the membrane fraction after Dounce homogenization. The majority of the S4GGnM-BSA-specific binding activity, 64% of the total, was found in the endothelial and Kupffer cells, while 36% was in hepatocytes.
We examined the ability of Triton X-100 to solubilize the binding
activity from membranes. Membranes were incubated with
125I-S4GGnM-BSA in the presence of increasing amounts of
Triton X-100 (Fig. 1). Membranes were either collected
by sedimentation in an Airfuge at 190,000 × g
(Beckman) in the absence of added PEG or by filtration on GF/C filters
following addition of PEG. 125I-S4GGnM-BSA was found in the
membrane pellet (Fig. 1, PEG) following addition of 0.05%
Triton but not 0.01%, 0.5%, or 1.0% Triton. In contrast,
125I-S4GGnM-BSA was bound in the presence of 0.05% Triton
as well as 0.5% and 1.0% Triton when complexes were collected by
precipitation with PEG (+PEG). At a Triton concentration of
0.05%, membrane vesicles become sufficiently permeable for the
S4GGnM-R to be accessible to the 125I-S4GGnM-BSA; however,
the S4GGnM-R is not solubilized at this Triton concentration and can be
sedimented in the absence of added PEG. At Triton concentrations above
0.5%, the S4GGnM-R is fully solubilized and requires PEG for
precipitation of complexes. Using both the sedimentation assay in the
presence of 0.05% Triton and the PEG precipitation assay in the
presence of 1% Triton, we determined that: 1) 3-fold more
125I-S4GGnM-BSA is bound at pH 7.5 than at pH 5.0 or below;
2) EDTA does not abolish binding of S4GGnM-BSA but does reduce it to
62% of that seen in the presence of 4 mM Ca2+;
and 3) fucoidin is a significantly more potent inhibitor of binding
than other sulfated or anionic polysaccharides such as hyaluronic acid,
heparin, chondroitin sulfate, and dextran sulfate. Thus, a binding
activity with the properties expected for the S4GGnM-R could be
detected in rat liver membranes and solubilized with Triton X-100. We
therefore developed the isolation scheme summarized in Table
I.
|
The S4GGnM-R was solubilized using 10% Triton X-100, concentrated by
precipitation with PEG 8000, and solubilized in 10% Triton X-100 prior
to incubation with WGA-Sepharose. The S4GGnM-R bound to WGA-Sepharose
and was eluted with 300 mM GlcNAc. The WGA-Sepharose eluate
containing the S4GGnM-R was incubated with bLH-Sepharose. S4GGnM-R that
bound to bLH-Sepharose was eluted by reducing the pH to 4.0 with
acetate buffer. When this material was examined by SDS-PAGE, a major
band was found to be present, which had an Mr of
180,000 (Fig. 2A). Additional proteins with
apparent molecular weights of 75-80,000, 60,000, 43,000, and 35,000 were also present. Following electrophoretic transfer to Immobilon-P, a
ligand blot was performed using 125I-S4GGnM-BSA, which
demonstrated that only the protein with an Mr of
180,000 was reactive (Fig. 2B). An NH2-terminal
sequence of LK(Y)S(Q)YQFLIYNE was obtained for the protein
with an Mr of 180,000, suggesting it was closely
related to the murine macrophage mannose receptor (Man-R), which has an
NH2-terminal sequence of LLDARQFLIYNE (23).
S4GGnM-R that had been eluted from bLH-Sepharose was incubated with
mannan-Sepharose or mannose-Sepharose. Neither the S4GGnM-BSA-specific binding activity nor the protein with an Mr of
180,000 was bound by immobilized mannan or mannose, whereas the other
proteins in the eluate from bLH-Sepharose were bound to either
mannan-Sepharose or mannose-Sepharose and removed. The
mannose-Sepharose unbound fraction was homogeneous and consisted of a
single band, migrating with an Mr of 180,000 when analyzed by SDS-PAGE (see Fig. 4), which was designated the
S4GGnM-R.
The NH2-terminal sequence of the S4GGnM-R suggested a close relationship to the macrophage Man-R. However, the inability of the S4GGnM-R to bind to either immobilized mannan or mannose, which are used for affinity-based purification of the Man-R from lung (24), placenta (25), and macrophages (26, 27), indicated the S4GGnM-R is distinct from the Man-R. We examined the potential relationship between the S4GGnM-R and the Man-R by isolating the macrophage Man-R from rat lung by affinity chromatography on Mannose-Sepharose as described by Lennartz et al. (24) for direct comparison with the S4GGnM-R isolated from rat liver. The Man-R isolated from lung by this procedure is homogeneous and has an Mr of 180,000 when examined by SDS-PAGE (see Fig. 4). As will be presented in greater detail below, the S4GGnM-R and Man-R have a number of features that indicate they are distinct and other features that indicate they are closely related. For example the following features indicate that the receptors differ. 1) The S4GGnM-R will bind to immobilized ligands containing terminal GalNAc-4-SO4 but not ligands with terminal Man, while the Man-R will bind to immobilized ligands containing terminal Man but not those containing terminal GalNAc-4-SO4. 2) The S4GGnM-R is able to bind soluble ligands terminating with S4GGnM as well as ligands with terminal Man or Fuc in the PEG precipitation assay, whereas the Man-R will bind ligands with terminal Man or Fuc but not those with terminal S4GGnM in the same assay. Features indicating the receptors are structurally related include the following observations. 1) The S4GGnM-R and the Man-R both react with 125I-Man-BSA in ligand blots. 2) Antibodies raised to either the purified S4GGnM-R from liver or the Man-R from lung react with either receptor in Western blots. 3) The receptors have similar peptide maps, and multiple peptides prepared from the S4GGnM-R have sequences that are identical to those of the murine macrophage Man-R (23).
The S4GGnM-R and the Man-R Provide Comparable Peptide MapsThe S4GGnM-R (200 µg) and the Man-R (150 µg) were
subjected to electrophoretic separation on 5% polyacrylamide gels and
electrophoretically transferred to Immobilon (Millipore) in CAPS
buffer. After staining with Ponceau Red, the regions containing the
transferred protein were excised for analysis. Peptides were released
by digestion with LysC or trypsin in the presence of reduced Triton
X-100 and separated by reverse phase chromatography. The separations
are shown in Fig. 3 for peptides released by LysC
digestion of the S4GGnM-R and the Man-R. The profiles were nearly
identical. Only peaks 57 and 75 of the S4GGnM-R were not also present
in the Man-R. Peaks 45, 62, and 91 of the S4GGnM-R appeared to be
identical to peaks 37, 54, and 84 of the Man-R, respectively. Peaks
that appeared to be identical and peaks that appeared to differ between the S4GGnM-R and the Man-R were analyzed. The results of these analyses
are summarized in Table II.
|
Sequence was obtained for 12 peptides from three different S4GGnM-R
preparations. Nine of the peptide sequences obtained from the S4GGnM-R
were identical to peptide sequences predicted to be present in the
macrophage Man-R (23). The sequences obtained originate from a number
of different regions and encompass the entire 1365 amino acids of the
Man-R extracellular domain (Table II). As expected from the similarity
of the peptide maps, the predominant peptides present in peaks 45, 62, and 91 of the S4GGnM-R were identical to those of peaks 37, 54, and 84 of the Man-R, respectively. Only in the case of peaks 57 and 78 of the
S4GGnM-R, which had no equivalent in the Man-R peptide profile, were
sequences obtained that could not be identified in the Man-R sequence.
These were minor components, however, and were identified within the asialoglycoprotein receptor (ASGP-R) sequence. Since small amounts of
terminal 1,4-linked GalNAc are present on bLH and would be recognized by the ASGP-R, the presence of low levels of ASGP-R in the
final preparation of the S4GGnM-R would not be unexpected. We did not,
however, detect ASGP-R in the S4GGnM-R preparation by Western blot
analysis using ASGP-R-specific antisera.
The peptide sequence analyses summarized in Table II, in conjunction with the nearly identical peptide maps obtained for the S4GGnM-R and the Man-R in Fig. 3, support the conclusion that the S4GGnM-R from liver is closely related to the macrophage Man-R isolated from rat lung. Furthermore, the S4GGnM-R and the Man-R both react with antisera raised to either receptor in Western blots (data not shown). Despite the strong evidence of a close structural relationship between the S4GGnM-R and the Man-R across their entire extracellular regions, we had evidence that indicated they would differ in specificity and ligand binding properties. We therefore examined their ligand binding properties in greater detail.
The S4GGnM-R and Man-R Differ in Their Ability to Bind to Immobilized GalNAc-4-SO4 and MannoseWhen purified
S4GGnM-R was incubated with Man-BSA conjugated to Sepharose (Fig.
4), >80% of the S4GGnM-R was present in the unbound
fraction when analyzed by SDS-PAGE and staining with Coomassie Blue
(panel A). Ligand blotting of the S4GGnM-R with
125I-S4GGnM-BSA (panel B) and
125I-Man-BSA (panel C) demonstrated that the
S4GGnM-R in the unbound fraction could react with soluble S4GGnM-BSA
and with soluble Man-BSA despite its inability to bind Man-BSA
conjugated to Sepharose. The Man-R, which had been isolated from lung
by affinity chromatography on mannose-Sepharose, did not bind to
bLH-Sepharose, which contains multiple oligosaccharides terminating
with S4GGnM (Fig. 5). Thus, the S4GGnM-R isolated from
liver and the Man-R isolated from lung differ in their ability to bind
to immobilized ligands containing terminal S4GGnM and Man. The S4GGnM-R
and the Man-R represent the major forms of these receptors in liver and
lung, respectively, since significant amounts of Man-R cannot be
isolated from liver by affinity chromatography on Man-Sepharose
following solubilization with Triton X-100 as described for isolation
of the Man-R from lung (24), nor can significant amounts of S4GGnM-R be
isolated from lung using the procedures we developed for isolation of
the S4GGnM-R from liver. This conclusion is supported by analysis of
the various steps during purification of the S4GGnM-R and the Man-R by
Western blotting with receptor-specific antisera (data not shown).
Comparison of Man-BSA Binding by the S4GGnM-R and the Man-R
The ability of the S4GGnM-R to bind to immobilized bLH but
not immobilized Man-BSA, and the ability of the Man-R to bind
immobilized Man-BSA but not immobilized bLH, suggested that the
S4GGnM-R and the Man-R would display different specificities for
soluble ligands when examined in the precipitation assay. Notably, the
S4GGnM-R binds both S4GGnM-BSA and Man-BSA in a
concentration-dependent manner as determined by
precipitation with PEG (Fig. 6A). In
contrast, the Man-R binds Man-BSA but does not bind S4GGnM-BSA using
the PEG precipitation assay (Fig. 6B). The inability of the
Man-R to precipitate S4GGnM-BSA is consistent with its inability to bind to immobilized bLH, which bears multiple Asn-linked
oligosaccharides terminating with the sequence S4GGnM (3). The ability
of the S4GGnM-R to precipitate soluble Man-BSA was not expected since the S4GGnM-R is not able to bind to immobilized ligands containing terminal mannose such as mannan-Sepharose, Man-Sepharose, and Man-BSA-Sepharose.
The S4GGnM-R and the Man-R both react with Man-BSA when examined by
ligand blotting with 125I-Man-BSA following SDS-PAGE (Fig.
7). When equal units of Man-BSA-specific binding
activity, as measured by the PEG precipitation assay, were examined by
ligand blotting with 125I-Man-BSA, the S4GGnM-R in
lane 3 was more intensely labeled than the Man-R in
lane 4. This suggested that even though 5-6-fold more
S4GGnM-R than Man-R was required to precipitate the same amount of
soluble Man-BSA (compare panels A and B of Fig.
6), this difference in binding capacity was not retained in the ligand blot following SDS-PAGE. In support of this conclusion, we found that
equal amounts of the S4GGnM-R and Man-R reacted with equal intensity
when examined by ligand blotting with 125I-Man-BSA
following SDS-PAGE (Fig. 7, lanes 1 and 2). The
difference in binding efficiency measured in the PEG precipitation
assay for the S4GGnM-R and the Man-R is not retained in ligand blots following separation by SDS-PAGE. This suggests that binding in the
soluble assay may reflect binding to a different site or in a different
manner than when the same receptor is probed with ligand following
SDS-PAGE. In light of the structural similarities between the S4GGnM-R
and the Man-R, it is notable that they bind Man-BSA with equal
intensity following SDS-PAGE even though they show marked differences
in Man-BSA binding in their native state.
A remarkable feature of the macrophage Man-R is its ability to bind
ligands with terminal Man, GlcNAc, Glc, and Fuc (28, 29). These same
monosaccharides can be utilized as inhibitors of binding by the Man-R.
We therefore compared inhibition of Man-BSA binding by monosaccharides
for both the S4GGnM-R and the Man-R. As reported by others (28, 30,
31), we found binding of Man-BSA by the Man-R in the PEG precipitation
assay is inhibited by Man, Fuc, Glc, and GlcNAc, whereas Gal inhibits
weakly and GalNAc not at all (Fig. 8). In contrast, low
concentrations of Man enhance Man-BSA binding by the S4GGnM-R.
Concentrations of Man as high as 200 mM are not inhibitory,
although they do reduce binding as compared with Man concentrations
ranging from 25-100 mM (Fig. 8). Fuc, Glc, and GlcNAc have
similar effects whereas GalNAc is without effect (Fig. 8). Gal, which
is a poor inhibitor of binding by the Man-R, enhances binding by the
S4GGnM-R at a concentration of 200 mM but not at 50 mM. Thus, the same monosaccharides affect binding by the
S4GGnM-R and the Man-R; however, they enhance binding by the S4GGnM-R
and inhibit binding by the Man-R. Even though the S4GGnM-R and the
Man-R both are able to bind Man-BSA, the properties of this binding
reaction for the native receptors differ dramatically with respect to
the effect of monosaccharides and the amount of receptor required to
precipitate a given amount of Man-BSA.
Man-BSA and S4GGnM-BSA Bind to the S4GGnM-R Independently
We
next determined if there was any relationship between the mannose- and
S4GGnM-specific binding sites of the S4GGnM-R. Like the Man-R, we found
that the S4GGnM-R is able to bind Fuc-BSA as well as Man-BSA (Fig.
9). Addition of excess Man-BSA inhibited binding of
125I-Man-BSA and 125I-Fuc-BSA by the S4GGnM-R.
Man-BSA had no effect on the binding of 125I-S4GGnM-BSA by
the S4GGnM-R (Fig. 9). Fuc-BSA was also able to inhibit binding of both
125I-Man-BSA and 125I-Fuc-BSA by the S4GGnM-R,
suggesting Man-BSA and Fuc-BSA compete for the same sites on the
receptor. In contrast excess S4GGnM-BSA inhibits binding of
125I-S4GGnM-BSA by the S4GGnM-R but not binding of either
125I-Man-BSA or 125I-Fuc-BSA (Fig. 9). The
addition of either 40 mM mannose or fucose enhanced binding
of Fuc-BSA and to an even greater extent than Man-BSA. At a
concentration of 200 mM, binding of Man-BSA and Fuc-BSA
were reduced as compared with that seen in the presence of 40 mM monosaccharide but not to the levels seen in the
complete absence of added monosaccharides. Neither mannose nor fucose
had any effect on S4GGnM-BSA binding at either concentration (Fig. 9).
Thus S4GGnM-BSA appears to bind to a site on the S4GGnM-R that is
distinct from and independent of the Man/Fuc-specific binding site.
Kinetics of Man-BSA and S4GGnM-BSA Binding
The kinetics of
binding of S4GGnM-BSA and Man-BSA in the presence and absence of GlcNAc
were assessed for both the S4GGnM-R and the Man-R. The saturation
curves obtained were analyzed using the Ligand program (32) as
summarized in Table III. At saturation, 0.06 mol of
S4GGnM-BSA/mol of S4GGnM-R was bound with an apparent Kd of 3.0 × 109 M.
The presence of 50 mM GlcNAc had no impact on the kinetics of S4GGnM-BSA binding. Man-BSA was bound by the S4GGnM-R with an
apparent Kd of 3.9 × 10
9
M and a mole ratio of 0.11 at saturation. In the presence
of 50 mM GlcNAc, the apparent Kd for
binding of Man-BSA by the S4GGnM-R was reduced to 1.7 × 10
8 M while the mole ratio at saturation
increased to 1.06. The apparent Kd for binding of
Man-BSA by the Man-R from lung was 3.1 × 10
9
M with a mole ratio of 0.77 at saturation. As with the
S4GGnM-R, 50 mM GlcNAc reduced the apparent
Kd to 1.5 × 10
8 M
and increased the mole ratio to 1.05 at saturation.
|
Thus, even though the S4GGnM-R is able to bind both S4GGnM-BSA and Man-BSA, the kinetics seen for binding of Man-BSA differ from those seen for binding of Man-BSA by the Man-R. Since both the apparent Kd and the Bmax for binding of Man-BSA by the S4GGnM-R are influenced by addition of monosaccharides such as GlcNAc, the effects of monosaccharides on binding by the S4GGnM-R must be considered complex and will require more detailed analysis to be understood. It would appear, however, that the S4GGnM-R is capable of binding and internalizing ligands with terminal GlcNAc, Man, or Fuc as well as those with terminal S4GGnM. Thus, it is not clear at present if the receptor responsible for clearance of glycoproteins bearing oligomannose type oligosaccharides or neoglycoproteins such as Man-BSA from the blood (33) is the S4GGnM-R or the Man-R.
We have isolated a protein that has the properties predicted for the S4GGnM-R in hepatic endothelial cells and Kupffer cells. The S4GGnM-R mediates the rapid clearance of glycoproteins bearing oligosaccharides with the terminal sequence S4GGnM, for example LH and TSH, from the circulation (12). We have proposed that this function is critical for the expression of hormone biologic activity in vivo (6-9, 34). Peptide maps, amino acid sequence of multiple peptides, and immune cross-reactivity indicate the S4GGnM-R is closely related to the previously characterized macrophage Man-R (23, 35). The Man-R is one member of a family of structurally related membrane proteins, which includes DEC-205 (36), the phospholipase A2 receptor (37, 38), and a "novel" type C lectin expressed in fetal but not adult liver (39). Each consists of a signal sequence that is cleaved, a cysteine-rich (Cys-R) domain, a domain with fibronectin type II repeats (FN-II), 8-10 type C carbohydrate recognition domains (CRD) separated by "stalks," a transmembrane domain, and a cytosolic domain. The S4GGnM-R peptides for which sequence was obtained (Table II) represent sequences that are identical to portions of the Cys-R domain, the FN-II domain, CRD 2, CRD 3, CRD 5, CRD 8, and Stalk 2 of the macrophage Man-R. If we assume that this similarity in structure is retained throughout the extracellular domain of the S4GGnM-R, it will consist of a Cys-R domain, a FN-II domain, 8 CRDs, 9 stalks, a transmembrane domain, and a cytosolic domain.
How closely related are the S4GGnM-R and the Man-R and what is the structural basis for the differences in their ligand specificities? The data we have obtained indicates that the structural relationship between the S4GGnM-R and the Man-R is extensive. Possible mechanisms that could result in two such closely related receptors include: 1) the existence of two genes encoding closely related proteins, 2) a post-transcriptional alteration in the mRNA sequence producing different forms of the receptor, and 3) a post-translational modification which alters the specificity of the receptor. Any differences in the structure of the S4GGnM-R and the Man-R must account for differences in the ability to bind ligands terminating with S4GGnM and for differences in the characteristics for binding of ligands containing terminal mannose or fucose.
The macrophage Man-R displays multiple specificities, being able to bind carbohydrate moieties terminating with Man, GlcNAc, or Fuc (28-30). CRDs 4-8 together account for the high affinity binding of ligands such as Man-BSA and mannan. CRD 4 plays a predominant role in binding and is the only CRD that is able to bind mannose containing ligands in the absence of other CRDs (28). Binding of Man-BSA and Fuc-BSA by the S4GGnM-R, like binding by the Man-R, is Ca2+- and pH-dependent. Furthermore, the same monosaccharides have an impact on binding by both receptors; however, in the case of the Man-R these monosaccharides are strictly inhibitory, whereas for the S4GGnM-R they have a complex effect resulting in enhanced rather than reduced binding. This suggests that one or more of the CRDs that mediate Man-BSA binding by the Man-R are altered in the S4GGnM-R. Based on peptide sequences from the S4GGnM-R the Cys-R region, FN-II region, CRD 2, and CRD 3 are present and likely have the same sequence as in the Man-R. The functional significance of the Cys-R region and FN-II region is not known for the Man-R or other members of this family, nor is there evidence of ligand binding by CRDs 1-3 of the Man-R (28). Should the S4GGnM-R prove to have the same overall structure as the Man-R throughout its extracellular domain, there would be an ample number of regions, which could potentially account for the independent binding of ligands terminating with S4GGnM and those terminating with either mannose or fucose.
The S4GGnM-R may be the first example of a carbohydrate-specific receptor that can bind unrelated oligosaccharide structures at independent sites. Since binding of S4GGnM-BSA does not require Ca2+, it is likely that the structural motif which accounts for binding of GalNAc-4-SO4-containing ligands differs from the Ca2+-dependent CRDs, which are characteristic of the Man binding sites. If distinct regions account for S4GGnM- and Man-specific binding by the S4GGnM-R, it would seem likely that structural differences between the S4GGnM-R and the Man-R would have to involve more that one region. An intriguing feature of the S4GGnM-R is that the same peptide bearing different carbohydrate moieties could be bound by the same receptor at different sites and with differing kinetics. This could result in differing kinetics of clearance from the circulation. For example, LH bearing high mannose type structures may be cleared more rapidly than LH bearing oligosaccharides terminating with S4GGnM. Should this be the case it would suggest that the precise rate of clearance as determined by the structure of the oligosaccharide is indeed critical for maintaining biologic activity in vivo.
A number of studies have suggested that a receptor with the same properties as the Man-R isolated from lung and placenta is present in liver endothelial cells and Kupffer cells (30, 40-42). The relationship of the hepatic receptor to the macrophage Man-R present in alveolar macrophages and other tissues may have to be re-evaluated in light of the current findings. It is not surprising that it has been difficult to purify the Man-R from liver using affinity chromatography on immobilized ligands containing terminal mannose, procedures that are effective for isolation of the Man-R from lung (24), macrophage lines (26), and placenta (25), in light of the properties of the S4GGnM-R. It now seems likely that the S4GGnM-R we have isolated from liver accounts for a major fraction of binding and internalization of ligands containing terminal Man, Fuc, or GlcNAc by hepatic endothelial cells and Kupffer cells (41-43). The characteristics of binding and internalization of Man-BSA by the S4GGnM-R are likely to differ from those encountered with the Man-R of alveolar macrophages.
Many issues remain to be addressed. How does the S4GGnM-R differ from the Man-R structurally? What region accounts for binding of S4GGnM? Is the region accounting for S4GGnM binding structurally related to the Ca2+-dependent C type lectin motif, or does it represent a new binding motif? Is expression of the S4GGnM-R, like that of the Man-R, highly regulated. Is expression regulated by estrogen? Is the S4GGnM-R expressed in other cells and in other tissues? These are some of the issues we will address in our future studies. The answers promise to reveal new insights about the biologic significance of oligosaccharides terminating with S4GGnM for the glycoprotein hormones as well as other glycoproteins, which are continually being added to the list family of glycoproteins bearing S4GGnM structures.
We thank William S. Lane and colleagues at the Harvard Microchemistry Facility for their helpful suggestions while performing the peptide maps and sequence analyses. We also thank Dr. P. D. Stahl, Washington University, St. Louis, MO for providing us with antibody raised to the Man-R and Dr. R. L. Hill, Duke University, University Medical Center, Durham, NC for providing samples of Fuc-BSA and antibody raised to the Man-R.