The distribution and localization of the fucose-binding lectin in rat tissues and the identification of a high affinity form of the mannose/N-acetylglucosamine-binding lectin in rat liver.

A small-scale affinity chromatographic procedure was developed to screen for the presence of fucose and mannose/N-acetylglucosamine-binding lectins in small amounts of rat tissues. Of all tissues examined, only the liver contained the fucose-binding lectin, whereas both liver and blood serum contained the mannose/N-acetylglucosamine lectin. By means of immunocytological methods using antibodies to hepatic lectins, the fucose lectin was shown to be uniquely present in Kupffer cells and absent in all other types of rat macrophages examined. The binding and uptake of different neoglycoproteins by nonparenchymal cell fractions of liver indicated that the fucose-binding lectin was either not responsible for the uptake or that more than one lectin was acting. With the identification of another lectin (Mr = 180,000) by the above screening procedure for hepatic lectins and the results of studies in the following paper (Haltiwanger, R.S., and Hill, R. L. (1986) J. Biol. Chem. 261, 7440-7444) two lectins appear to be involved. A small amount of the hepatic mannose/N-acetylglucosamine lectin was found by the above screening procedure to have a higher affinity for L-fucosyl-bovine serum albumin-Sepharose than the majority of the lectin in hepatocytes. This lectin, called the high affinity form, was purified and its properties examined. On a weight basis the high affinity form bound 7-12 times more ligand than the normal form. Its Ka for L-fucosyl-bovine serum albumin was 2.3 X 10(9) M-1 compared to 3.5 X 10(8) M-1 for the normal form. Moreover, the concentrations of monosaccharides required to inhibit the high affinity form were about 3 times less than those required to inhibit binding of the normal form. The two forms, however, have identical molecular weights (32,000) under reducing and nonreducing conditions, bind anti-lectin antibodies in the same way, and give identical peptide maps after V-8 protease digestion. The structural basis for the different binding affinities of the two forms remains unknown.

The preceding papers describe the purification, some structural properties (l), and the binding specificity (2) of a lectin from rat liver with a high affinity for ligands containing nonreducing terminal L-fucose or D-galactose. It differs in its structural properties from the other known lectins of liver, including those with a binding specificity for galactose (3,4) and mannose/N-acetylglucosamine (5, 6). Although it binds L-fucose and D-galactose derivatives strongly, its affinity for N-acetylglucosamine is very low. The mannose/N-acetylglucosamine lectin, however, binds L-fucose in addition to mannose and N-acetylglucosamine but has a low affinity for galactose. In contrast, the galactose lectin has a high affinity for galactose and a low affinity for L-fucose, mannose, and Nacetylglucosamine.
In order to understand the possible functions of these hepatic lectins, including their participation in the uptake of circulating glycoproteins from blood (7,8), the tissue distribution of the fucose lectin and its cellular localization in the liver were studied. The results of these studies, reported here, show that the fucose lectin is present in hepatic Kupffer cells, in contrast to the galactose (7) and N-acetylglucosamine lectins (9), which are in hepatocytes. Moreover, since no other rat tissue or type of macrophage contains the fucose lectin, it appears to be a protein unique to Kupffer cells.
Also reported here is the isolation and characterization of a form of the mannose/N-acetylglucosamine lectin that has a higher affinity for its ligands than the normal form isolated from liver extracts. This lectin, called the high affinity form, was discovered during the course of purification of the fucose lectin (1) and was easily identified in a screening procedure developed for detecting lectins in small amounts of rat tissue. The screening procedure also disclosed a hepatic protein (Mr = 180,000) with a high affinity for Fuc-BSA.' Since the pattern of binding of neoglycoproteins to nonparenchymal liver cells could not be explained by the binding specificity of the fucose lectin, another yet unidentified lectin, perhaps corresponding to the protein with a M, = 180,000, was suggested to be involved. As shown in the following paper (IO), which describes the purification and partial characterization of an alveolar macrophage lectin (Mr = 180,000) with a high affinity for N-acetylglucosamine and L-fucose, this proved to be the case.

EXPERIMENTAL PROCEDURES
Materials-Rabbit antibodies to the rat liver fucose lectin and the mannose/N-acetylglucosamine lectin were prepared as described pre-Neoglycoproteins are abbreviated with the standard symbols for monosaccharides and bovine serum albumin. Thus, Fuc-BSA is Lfucosyl-bovine serum albumin. Unless otherwise designated all monosaccharides used here were of the D-configuration except fucose which was of the L-configuration. The other abbreviation used is: HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid.

7433
viously (1). Antiserum to the rat liver galactose lectin was prepared in the same manner except that 76 pg of the purified lectin was injected followed by two boosts of 75 pg each. The IgG in antisera was isolated on Protein A-Sepharose 4B columns (11). F(ab')n fragments were prepared as described (12) by pepsin digestion. Goat antirabbit IgG was labeled with fluorescein isothiocyanate by published procedures (13).
The normal form of the mannose/N-acetylglucosamine lectin was prepared as described earlier (1). The high affinity form was prepared from the material obtained in step 5 of the normal purification procedure (1) in which the lectin was adsorbed on anti-lectin IgG-Sepharose. The adsorbent was suspended in 1-2 volumes of 0.2 M glycine-HCI, pH 2.2, for 60 min at 23 "C. The eluate was adjusted to pH 7.8, made 0.01 M in calcium chloride and purified on a column (1.2 X 2.5 cm) of Fuc-BSA-Sepharose as described in step 4 of the purification procedure (1). The anti-lectin IgG-Sepharose was regenerated by exhaustive washing with 0.01 M Tris-C1, pH 7.8, containing 0.01 M sodium chloride and 0.02 M sodium azide, and stored at 4 "C. All other materials were obtained as described earlier (1,2).
Distribution of Lectins in Rat Tissues-The alternate procedure (1) for preparation of the fucose lectin was scaled down and used to analyze the lectins in different tissues. Tissue (0.5 g) was homogenized in 3-4 ml of extraction buffer exactly as in step 1 of the purification procedure. The homogenate was then carried through step 2 (precipitation with polyethylene glycol) exactly as described. The material from step 2 was applied to 1 ml of Fuc-BSA-Sepharose (10 mg of Fuc-BSA/ml) in a column (0.4 X 2 cm) equilibrated with buffer A (1). The flow rate of the column was reduced to t l ml/min to prevent fissure formation when all buffer flowed through. After washing with 10 ml of Buffer A the column was eluted with 5 ml of Buffer B. The column was regenerated by washing with 5 ml of 0.05 M sodium acetate, pH 4.0, and 5 ml of Buffer B containing 0.05 M N-acetylglucosamine. The Buffer B eluates (5 ml) were made 0.05 M in Nacetylglucosamine and 0.01 M in calcium chloride and applied to the regenerated column. The eluate (5 ml) contained the normal form of the mannose/N-acetylglucosamine lectin and was combined with the first 2 ml of the eluate obtained by further elution of the column with 10 ml of Buffer A containing 0.05 M N-acetylglucosamine. After elution with 10 ml of Buffer A, the columns were eluted with 4 ml of Buffer A containing 0.04 M galactose followed by 4 ml of Buffer B. The Buffer B eluate contained the high affinity form of the mannose/ N-acetylglucosamine lectin. The columns were regenerated as above 95% ethanol and stored at -20 "C for 2 h. The precipitates were for reuse. The lectins in each fraction were mixed with 4 volumes of collected by centrifugation at 27,000 X g for 30 min, dissolved in 50 pl of sample buffer for gel electrophoresis, and analyzed on 10% polyacrylamide gels in sodium dodecyl sulfate (14). The proteins were detected by staining the gels with silver nitrate (15).
Immunofluorescent Staining of Frozen Sections-Freshly isolated rat tissue was immersed in OCT compound and quickly frozen on solid COz. Frozen sections (8 pm) were cut on a cryostat and fixed onto glass slides with absolute ethanol at -20 "C for 20 min. The phosphate, pH 7.4, 0.15 M NaCl, 0.02% NaN3 (PBS). Each section sections were washed three times for 10 min each in 10 mM sodium was then covered with a drop of PBS plus 1% BSA containing 50 pg/ ml of the appropriate IgG and incubated for 2 h at 25 "C. After washing as before, the sections were incubated with a drop of PBS plus 1% BSA and 50 pg/ml fluorescein isothiocyanate-labeled goat anti-rabbit IgG for 2 h at 25 "C. The sections were then rinsed for a final time, mounted, and viewed with a Zeiss photomicroscope I11 with epifluorescence. Pictures were taken with Ektachrome 400 film which were push. processed to ASA 1600.
Western Blots-Western blots were performed essentially as described earlier (16, 17) except that 9% electrophoretic gels were used and lZ5I-goat anti-rabbit IgG was used rather than '251-protein A.
Tissues or cells were homogenized in 10 volumes of hot 65 mM Tris .
Any undissolved material was removed by centrifugation. Bromphenol blue and 2-mercaptoethanol were added to the samples to give final concentrations of 0.00125 and 5%, respectively, before heating each sample to 100 "C for 5 min.
Pronase Digestion of Liver-Liver slices were hydrolyzed with Pronase as described (18). After filtering the digested liver through 100-pm nylon mesh, the nonparenchymal cells were centrifuged at 250 X g for 10 min and washed two times with Eagle's modified essential medium containing 10% fetal calf serum, 2.5 D M CaCl,, and 20 mM HEPES, pH 7.4. Typically 1-5 X 10' nonparenchymal cells were recovered per rat.
Collagenase Digestion of Liver-Collagenase digestion of rat liver was performed by recirculating perfusion (19, 20). Type I11 collagenase from Worthington was found to have the least adverse affects on the lectin and was used for routine perfusions. After digestion, the disrupted liver was filtered through 100-pm nylon mesh and the parenchymal cells removed by centrifugation at 10 X g for 5 min. The nonparenchymal cells were collected by centrifuging at 250 X g for 10 min and washed three times with Hanks' balanced salt solution containing 0.0005% deoxyribonuclease I. The nonparenchymal cells were enriched in Kupffer cells with Percoll gradient centrifugation as described (21). Approximately 5-10 X lo7 cells were layered onto each of three 33-ml gradients and separated as described. Cells in the gradient corresponding to densities of between 1.045 and 1.062 g/ml were collected, diluted &fold with Hanks' solution and centrifuged at 250 X g. These cells were then resuspended in Hanks' solution and used immediately. Typically 3-14 X lo7 cells were isolated per rat with greater than 90% viability as judged by exclusion of trypan blue. In a routine preparation, 38% of the cells were peroxidase positive as determined by the method of Page and Garvey (22).
Uptake of Colloidal Iron by Kupffer Cells-Colloidal iron labeling of Kupffer cells was performed as described earlier (23). The liver turned gray-black after injection of the colloidal iron into the inferior vena cava. After 10 min to allow the iron particles to be phagocytized, the vena cava was cannulated and the liver was perfused in situ with phosphate-buffered saline (no azide) at a rate of 12.5 ml/min for about 5 min. The liver blanched after about 1 min but remained a grayish-green color due to the phagocytized iron particles. The liver was then removed and processed for immunofluorescence as described above.
Binding of 125Z-F(ab')2 Fragments of Preimmune and Immune Antifucose Binding Lectin ZgG to Isolated Nonparenchymal Cells-Cells isolated by Pronase digestion were incubated for 1-2 h at 37°C with shaking to allow replenishment of the lectin on the surface of cells.
The cells (100 pl) were chilled to 4 "C and incubated with 100 ng of 1251-F(ab')z fragments of both preimmune and immune anti-fucose lectin IgG in 100 pl of Eagle's media, 10% fetal calf serum, 2.5 mM CaC12, and 20 mM HEPES, pH 7.4. Both the preimmune and immune IgG were iodinated to the same specific activity of about 8.8 pCi/pg. After 2 h, 0.5 ml of Eagle's medium, 10% fetal calf serum, 2.5 mM CaC12, and 20 mM HEPES, pH 7.4, were added and the cells were collected by filtration on Whatman GF/C glass fiber filters previously soaked in the Eagle's media. The filters were rinsed with 15 ml of the same solution and counted in a y counter. Nonspecific binding to the fiiters in the absence of cells was less than 1% of the total counts added to each assay.
Binding of 1251-Neoglycoproteins to Nonparenchymal Cells at 37°C-Nonparenchymal cells freshly isolated by collagenase perfusion as described above were incubated at 37°C in Eagle's medium, 10% fetal calf serum, 2.5 mM CaC12, and 20 mM HEPES, pH 7.4, a t a concentration of 4 X lo6 cells/ml with 1 ng/ml of the appropriate '%Ineoglycoprotein (24). Aliquots (0.25 ml; lo6 cells) were removed at various times and layered on top of 0.15 ml of silicone oil (4 parts of DC550 to 1 part of light mineral oil). The cells were centrifuged for 30 s in an Eppendorf microcentrifuge to separate the cells from unbound ligand. The pellet in the tubes was removed by slicing through the oil layer and counted in a y counter. Blank values were obtained by addition of 1000-fold excess of the appropriate neoglycoprotein to the incubation mixture. Blank values were typically less than 0.1% of the total counts added to each assay.

RESULTS
Distribution of the Fucose Lectin in Rat Tissues-Two methods for purification of the fucose lectin were described earlier (1). One, the so-called "alternate procedure," separated the lectin from other lectins by affinity chromatography on FUC-BSA-agarose adsorbent in the presence and absence of specific monosaccharides, as indicated diagrammatically in Fig.  1. The SDS-gel electrophoretic patterns of the various fractions obtained from this procedure are shown in Fig. 2. Thus, fucose lectin is specifically eluted in pure form with D-galactose-containing buffers at step 5 ( Fig. 1). Most of the mannose/N-acetylglucosamine lectin is unretarded on the column in step 4 ( Fig. 1) and appears along with other contaminating proteins, but another fraction of this same lectin is tightly bound to the Fuc-BSA adsorbent and is obtained in step 6. This procedure for separating the lectins was scaled down so that a Triton X-100 extract of 0.5 g of rat liver was fractionated on a 0.4 x 2-cm column of Fuc-BSA adsorbent. Each of the fractions was then analyzed by gel electrophoresis in sodium dodecyl sulfate, and the protein species corresponding to each lectin detected with silver nitrate, as shown in Fig. 2. Silver nitrate detects as little as 10-20 ng of lectin; thus, this  Fig. 1. A, protein standards (0.2 pg) of human transferrin, bovine serum albumin, ovalbumin, and trypsinogen. B, pure fucose lectin. C, the unadsorbed proteins at step 4 (Fig. 1). D, the proteins (fucose lectin) eluted from Fuc-BSA-Sepharose at step 5 ( Fig. 1) with 0.04 M galactose. E, the proteins in the eluate at step 6 ( Fig. 1) including the high affinity form of the mannose/N-acetylglucosamine lectin (Mr = 32,000) and a high molecular weight (M, = 180,000) protein, presumably another lectin. F, the high molecular weight (Mr = 180,000) protein ( L a n e E ) after removal of the mannose/N-acetylglucosamine lectin on anti-lectin IgG-Sepharose. Silver nitrate was used to detect the proteins.
ading Lectin 7435 method is more sensitive than detection of the lectin by a binding assay.
In view of the sensitivity of the above procedure for detecting the fucose lectin, it was used to determine the distribution of the lectin in rat tissues. Of all the tissues examined, including kidney, heart, lung, brain, testes, serum, spleen, pancreas, thymus, and skeletal muscle, only liver contained the fucose lectin. The yield of lectin was about 800 ng/g of liver.
In other studies, samples (0.5 g) of mouse and human liver were analyzed (Fig. l), and each contained silver nitrate reactive species in the fractions that contain fucose lectin. The molecular weights of these species, however, did not correspond exactly to those of rat fucose lectin (rat, 88,000 and 77,000; mouse, 95,000 and 90,000; human, 25,000,45,000, and 60,000).
Localization of the Fucose Lectin in Hepatic Cells-Antibodies to the fucose lectin were used to determine which type of hepatic cells contained the lectin. Antibodies to the galactose and the mannose/N-acetylglucosamine lectins of hepatocytes were also used as controls. Fig. 3 shows Western blots of the three lectins with their homologous antibodies and indicates the high specificity of the antibodies employed. Fig. 4 shows frozen sections of rat liver treated initially with F(ab')2 fragments of either the anti-fucose lectin IgG or the anti-galactose lectin IgG and then with fluorescein-labeled F(ab'), fragments of goat anti-rabbit IgG. When the same experiments were performed with F(ab')z fragments from anti-mannose/N-acetylglucosamine IgG, no fluorescence was observed, in accord with the observation that this lectin is not bound to membranes (9). The pattern of staining by the galactose and fucose lectins is quite different. Since the galactose lectin is in hepatocytes, it would appear that the fucose lectin is in another cell type. This was confirmed in other experiments in which colloidal iron was injected intravenously into rats and frozen sections of the liver examined as shown in Fig. 5. Iron is specifically taken up by Kupffer cells and not by hepatocytes nor endothelial cells (23). Since the fluorescent cells are coincident with those containing iron, it is concluded that the fucose lectin is in Kupffer cells.
Further studies were performed to determine if macro-

FIG. 5. Frozen sections of rat liver from rats injected with colloidal iron particles.
After infusion of a rat with colloidal iron, the liver was sectioned and the frozen sections reacted with fucose lectin antibodies as in Fig. 4. The sections were examined by fluorescence (left) or phase contrast (right) microscopy of the same field to reveal the labeled antibody and colloidal iron particles, respectively. phages other than Kupffer cells contained the fucose lectin. Accordingly, lung, spleen, peritoneal macrophages, and blood monocytes were examined with anti-lectin F(ab')z and fluorescein-labeled anti-rabbit F(ab')z as described above, but no fluorescence was observed. These tissues and cells were also extracted with sodium dodecyl sulfate and the extracts submitted to Western blotting. Fig. 6 shows that only Kupffer cells contained the fucose lectin. A protein with a lower molecular weight than the lectin cross-reacted with the antibody and was found in all macrophage preparations.
Further studies were performed to determine whether the fucose lectin was on the external surface of Kupffer cells, as would be expected if it is involved in receptor-mediated endocytosis. Nonparenchymal cells freshly isolated from collagenase-perfused rat livers had little or no lectin on their surfaces as judged by binding with F(ab')z fragments of antilectin IgG. It was found, however, that the lectin was particularly susceptible to hydrolysis by different collagenase preparations as judged by reactions of thin sections of liver with the enzyme and staining for lectin as in Figs. 4 and 5. Nonparenchymal cells from Pronase-digested liver slices also contained little lectin as judged by binding of labeled F(ab'), fragments. But if the cells were incubated 1-2 h in tissue culture medium at 37 "C, then lectin appeared as judged by binding of labeled antibody, as shown in Fig. 7.
Binding and Uptake Neoglycoprotein by Nonparenchymal Cells from Liver- Fig. 8 shows that freshly isolated nonparenchymal cells regained their ability in time to bind to Fuc-BSA, Gal-BSA and GlcNAc-BSA when incubated in tissue culture medium at 37°C. When cells were incubated in me-  dium for 1 h at 37 "C in the absence of neoglycoproteins and their binding to the neoglycoproteins examined at 4 "C, only Fuc-BSA and Gal-BSA bound, although 20-25 times less of each was bound per lo6 cells than at 37 "C. GlcNAc-BSA, however, did not bind at 4 "C, suggesting that more than one lectin may be involved in binding the neoglycoproteins. Table I lists the amounts of '251-labeled Fuc-BSA, GlcNAc-BSA, and Gal-BSA bound to the cells and the percentage of inhibition of binding by a 1000-fold excess of the unlabeled neoglycoprotein. In accord with the above results (Fig. 8) these data show that the pattern of inhibition is also not that predicted by the binding specificity of the fucose lectin. This suggests that the fucose lectin does not mediate binding and uptake or that another lectin is also acting.
Characterization of a High Binding Affinity MannoselN-Acetylglucosamine Lectin-A small fraction of the mannose/ N-acetylglucosamine lectin had a high binding affinity for Fuc-BSA in both procedures (1) for purification of the fucose lectin. Indeed this was the major protein to contaminate the fucose lectin. Its high affinity is illustrated in Fig. 2, which shows that it is the last species to elute from the affinity adsorbent, and rather drastic conditions, removal of Ca2+, are required for elution. It was ordinarily removed by adsorption on anti-lectin-IgG-agarose adsorbents (1) and was obtained during preparation of the fucose lectin by elution of the antilectin IgG-adsorbent at pH 2.2 (see "Experimental Procedures"). Thus, its properties could be examined and compared with those of the mannose/N-acetylglucosamine lectin that did not have a high affinity for fucosyl-BSA adsorbents. Fig. 9 shows the binding of lZ5I-Fuc-BSA by the two forms of the lectin. On a weight basis the high affinity lectin bound 7-12 times more ligand than the normal species. Moreover, binding by a mixture of the two forms was the sum of the two separately. Fig. 10 shows Scatchard plots (25) of ligand binding for the two species. The high affinity form had an apparent K. = 2.3 X lo9 "' with 2.7 X lo-' sites/monomer (MI = 32,000), whereas the normal form had an apparent K, = 3.5

TABLE I
The binding and uptake of neoglycoproteins by nonparenchymnl cells of liver Freshly prepared cells were incubated a t 37 "C with the lZ5I-labeled neoglycoproteins in the presence or absence of a 1000-fold excess of unlabeled neoglycoprotein and the binding of labeled neoglycoprotein measured after 1 h as described under "Experimental Procedures.   Table I1 lists the concentrations for four monosaccharides required to inhibit by 50% the binding of '251-Fuc-BSA to both forms of the lectin. Clearly, 2.8-3.6 times lower concentrations of monosaccharide were necessary for 50% inhibition of the higher affinity form. The binding specificity, however, was the same for both forms.
The two forms of lectin appear to have indistinguishable structures. Thus, the high affinity and normal forms are readily adsorbed to anti-lectin IgG-agarose columns (I). In addition, both show the same protein species on gel electrophoresis in sodium dodecyl sulfate either in the presence or absence of reducing agent. Only a single species (MI = 32,000) is observed in reducing agent although species corresponding to dimers and trimers of 32,000 appear in the absence of reducing agent. Finally, peptide maps of V-protease digests of each species were indistinguishable.
Another Hepatic Lectin- Fig. 2 shows that another protein species with a M, = 180,000 was observed on gel electrophoresis in sodium dodecyl sulfate of the proteins obtained in the eluate at step 6 ( Fig. 1) of the small scale analysis of liver extracts. The high affinity form of the mannose/N-acetylglucosamine lectin could be removed from this fraction by adsorption on anti-lectin-IgG-Sepharose leaving the M, 180,000 species in pure form. Insufficient material, however, was obtained to determine its properties. On the basis of the results presented in the following paper (lo), it is now appar-    1 (1). Nonspecific binding was measured in the absence of lectin and subtracted from the total binding. ent that this species is likely another lectin derived from alveolar macrophages or nonparenchymal liver cells.

DISCUSSION
The studies reported here show clearly that the fucose lectin is unique to Kupffer cells, the stationary macrophages in the sinusoidal spaces of the liver. The immunocytological methods used for cellular localization of the lectin showed that the galactose lectin is only in hepatocytes, in accord with earlier reports (7). The mannose/N-acetylglucosamine lectin, however, could not be detected by the immunocytological methods, in agreement with the report that this hepatocyte lectin is not bound to membranes (9).
The fucose lectin appears to be at least partly on the external surface of Kupffer cells, since it cannot be detected in nonparenchymal cells freshly isolated from collagenaseperfused livers but is detectable after the cells have been incubated in tissue culture medium for 1-2 h. Its surface location suggests that it is bound to plasma membranes, in accord with the fact that it cannot be solubilized from liver homogenates except on extraction with detergents such as Triton X-100 (1).
The fucose lectin is also found exclusively in Kupffer cells, since it could not be detected in other rat macrophages or tissues by either a small-scale screening procedure that detects nanogram amounts of the lectin or by immunological methods. Thus, it may well prove to be useful in the future as a protein marker of Kupffer cells.
It seems reasonable to assume that the fucose lectin serves in the receptor-mediated endocytosis of glycoconjugates by Kupffer cells. The studies reported here show that nonparenchymal hepatic cells do indeed bind and take up neoglycoproteins, in accord with this suggestion. But the specificity of binding and uptake is not in accord with the binding specificity of the fucose lectin. Thus, either the lectin is not involved or another lectin is also acting in endocytosis of glycoconjugates by nonparenchymal cells. The results presented in the following paper indicate that the latter is likely the case, since a lectin has been identified in alveolar macrophages that is very similar in molecular weight to a protein (lectin) also identified in liver by the small sale screening procedure for hepatic lectins used here (Fig. 2). More importantly, antibodies to the alveolar lectin react with nonparenchymal cells in immunocytological studies (10). These results suggest that unlike alveolar macrophages, which contain only the M, 180,000 lectin, Kupffer cells contain the fucose lectin and perhaps the same lectin found in alveolar macrophages. Others have recently reported the occurrence of the M, 180,000 lectin in alveolar macrophages (26).
The studies reported here indicate that at least three hepatic lectins participate in the uptake of glycoconjugates from blood of mammals, uiz. the galactose lectin of hepatocytes and two lectins in nonparenchymal cells. One of the two nonparenchymal cell lectins is the fucose lectin found exclusively in Kupffer cells and the lectin first identified in alveolar macrophages that is also present in either Kupffer cells or endothelial cells, or both. Based on the different binding specificities of the three lectins, a wide variety of glycoconjugates with different nonreducing terminal monosaccharides would be readily removed from the circulation by the liver, in accord with earlier observations (7, 8). It should be emphasized, however, that the mannose/N-acetylglucosamine lectin of mammalian hepatocytes, which is not on the cell surface (9), is likely not involved in clearance of glycoconjugates from blood. In view of its similar binding specificity to the alveolar lectin, it has been assumed that it was present in macrophages and involved in uptake (5,6,27).
Finally, a high affinity form of the mannose/N-acetylglucosamine lectin was detected by the affinity chromatographic procedure used here to screen small amounts of tissue for the fucose lectin. This form is structurally indistinguishable from the normal form and represents as little as 1 4 % of the total lectin isolated. But it clearly has a higher affinity for its ligands, with a K, about 6 times the normal form. The structural basis for the high affinity form and its function remain unknown and will require further study.