Detection of novel carbohydrate binding activity of interleukin-1.

Tamm-Horsfall glycoprotein (THGP) and the oligosaccharide fraction liberated from THGP by hydrazinolysis inhibited tetanus toxoid-induced T cell proliferation. Intact THGP showed approximately 100-fold more inhibitory activity than the free oligosaccharides. After fractionating the oligosaccharides by anion-exchange column chromatography, the inhibitory activity could be detected in a sialidase-resistant acidic oligosaccharide fraction (fraction AR). The inhibitory activity of fraction AR was not observed when the fraction was added to the T cell culture medium 24 h after the addition of tetanus toxoid. Increased concentration of interleukin (IL) 1beta and decreased concentration of IL-2 were observed in the T cell culture medium after the addition of fraction AR. The oligosaccharides in fraction AR also inhibited the growth of an IL-1-dependent cell line, D10-G4. These results strongly suggested that the oligosaccharides in fraction AR bind to IL-1beta and suppress its cytokine activity. IL-1beta actually bound to the fraction AR immobilized on an amino-bonded thin layer plate. Fractionation of the oligosaccharides indicated that only oligosaccharides containing an N-acetylgalactosamine residue and a sulfate residue bound specifically to IL-1beta. Removal of either the sulfate residue or the N-acetylgalactosamine residue from the oligosaccharides abolished both the proliferation-inhibition and IL-1beta binding activities. Since IL-1beta did not bind to thyroid-stimulating hormone, which has the sulfate group at C-4 of the N-acetylgalactosamine residue in its N-linked sugar chains, the binding of IL-1beta toward oligosaccharides in fraction AR was considered to be highly specific.

Tamm-Horsfall glycoprotein (THGP) and the oligosaccharide fraction liberated from THGP by hydrazinolysis inhibited tetanus toxoid-induced T cell proliferation. Intact THGP showed approximately 100-fold more inhibitory activity than the free oligosaccharides. After fractionating the oligosaccharides by anion-exchange column chromatography, the inhibitory activity could be detected in a sialidase-resistant acidic oligosaccharide fraction (fraction AR). The inhibitory activity of fraction AR was not observed when the fraction was added to the T cell culture medium 24 h after the addition of tetanus toxoid. Increased concentration of interleukin (IL) 1␤ and decreased concentration of IL-2 were observed in the T cell culture medium after the addition of fraction AR. The oligosaccharides in fraction AR also inhibited the growth of an IL-1-dependent cell line, D10-G4. These results strongly suggested that the oligosaccharides in fraction AR bind to IL-1␤ and suppress its cytokine activity. IL-1␤ actually bound to the fraction AR immobilized on an amino-bonded thin layer plate. Fractionation of the oligosaccharides indicated that only oligosaccharides containing an N-acetylgalactosamine residue and a sulfate residue bound specifically to IL-1␤. Removal of either the sulfate residue or the N-acetylgalactosamine residue from the oligosaccharides abolished both the proliferation-inhibition and IL-1␤ binding activities. Since IL-1␤ did not bind to thyroid-stimulating hormone, which has the sulfate group at C-4 of the N-acetylgalactosamine residue in its N-linked sugar chains, the binding of IL-1␤ toward oligosaccharides in fraction AR was considered to be highly specific.
Most eukaryotic proteins occur as glycoproteins. The carbohydrate moieties of glycoproteins play important roles not only in modulation of protein properties such as stability and biological activities but also in various molecular recognition processes, including initial reaction in bacterial and viral infections, cell adhesion in inflammation and metastasis, differentiation, development, regulation, and many other intercellular communication events (1)(2)(3). Understanding the molecular mechanisms of carbohydrate recognition is therefore important for the biology of multicellular organisms. Carbohy-drate-binding proteins (lectins) are widely distributed in eukaryotic cells and mediate in many specific biological functions including intercellular recognition, protein trafficking, and primitive defense reactions (4). We will describe here the novel lectin-like property of a cytokine.
Tamm-Horsfall glycoprotein (THGP) 1 is a glycoprotein produced by kidney and contains approximately 30% carbohydrate. Serafini-Cessi's group (5)(6)(7) reported that THGP works as a potential suppressive agent of both the lymphocyte proliferation induced by phytohemagglutinin-L 4 treatment and the mixed lymphocyte reaction. They also suggested that the inhibitory activity resides in the sugar moiety of the glycoprotein (6, 7). Muchmore and co-workers (8 -10) found an immunosuppressive glycoprotein in the urine of pregnant women, and named it uromodulin. It inhibited T cell proliferation induced by tetanus toxoid as well as T lymphocyte proliferation induced by interleukin 1 (IL-1). They also suggested that the carbohydrate portion of uromodulin played a fundamental role in its inhibitory activity (11). That uromodulin is identical to THGP was confirmed by amino acid sequencing (12). Muchmore and Decker (13) proposed that uromodulin inhibits T cell proliferation via binding to IL-1␤ through its carbohydrate moieties and inactivates a mediator. However, Moonen et al. (14) rejected this explanation, because uromodulin interacted only with the denatured IL-1␤ adsorbed to the plastic plate but not with the native soluble IL-1␤. Therefore, the precise immunosuppressive mechanism of uromodulin has not yet been clarified.
We found that the oligosaccharide fraction, obtained by hydrazinolysis of THGP followed by N-acetylation, inhibits the T cell proliferation induced by tetanus toxoid. This paper reports the partial structural characterization of the oligosaccharides with this inhibitory activity.
Cell Lines-IL-2-dependent mouse T cell line, CTLL cells, was obtained from Dr. Takiguchi (Institute of Medical Science, University of Tokyo) and cultured in RPMI 1640 medium (ICN Biomedicals Ltd.) containing 10% fetal calf serum (FCS) (Cell Culture Laboratories, Cleveland, OH) and 2 units/ml human rIL-2 (kindly provided by Shionogi Co., Osaka, Japan) in 5% CO 2 at 37°C. IL-1␤-dependent T cell clone, D10-G4 (17,18), was obtained from Dr. Kakiuchi (Institute of Medical Science, University of Tokyo) and cultured in RPMI 1640 medium containing 10% FCS and 5% Con A supernatant, which was derived from the culture medium of rat splenic cells treated with Con A.
Analytical Methods-Anion-exchange column chromatography was performed using a fast protein liquid chromatography apparatus (Pharmacia Biotech) equipped with a Mono-Q HR5/5 column. Elution was programmed as follows: 5 mM sodium acetate, pH 4.0, for 10 min and then 500 mM sodium acetate, pH 4.0, for 20 min, at a flow rate of 1 ml/min at room temperature. Bio-Gel P-2 column chromatography using distilled water was performed to remove the salt or the monosaccharide from oligosaccharide fraction. Descending paper chromatography was performed using a solvent (1-butanol:ethanol:water ϭ 4:1:1, v/v). Affinity chromatography on W. floribunda agglutinin-agarose was performed as follows. The sample dissolved in 100 l of 10 mM phosphate-buffered saline (PBS), pH 7.4, was applied to a W. floribunda agglutinin-agarose column (2 ml) and eluted with 15 ml of the same buffer. Oligosaccharides bound to the column were eluted with 8 ml of the buffer containing 100 mM N-acetylgalactosamine. After elution, the oligosaccharide fraction was freed from N-acetylgalactosamine by passing through a Bio-Gel P-2 column. Mild methanolysis was performed according to the procedure described previously (19).
Glycosidase Digestion-Oligosaccharides were incubated with the following enzyme solutions: (i) A. ureafaciens sialidase (200 milliunits) in 50 l of 0.1 M acetate buffer, pH 5.0; (ii) jack bean ␤-N-acetylhexosaminidase (4 units) in 50 l of 0.3 M citrate phosphate buffer, pH 5.0. One drop of toluene was added to the reaction mixtures to inhibit bacterial growth during incubation. After being incubated for 48 h at 37°C, enzymes were inactivated by heating the reaction mixture in a boiling water bath for 3 min.
Preparation of THGP and Release of Its N-Linked Sugar Chains as Oligosaccharides-THGP was isolated from urine of a healthy adult male according to the procedure described previously (20). The THGP, thus obtained, gave a single band under both reducing and nonreducing conditions stained with Coomassie Brilliant Blue after being subjected to 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (data not shown). For the inhibitory assay, purified THGP solution was filtered to free from bacteria. Thoroughly dried THGP (50 mg), thus obtained, was subjected to 10 h of hydrazinolysis followed by N-acetylation as described previously (21). One-twentieth of the oligosaccharide fraction thus obtained was reduced with NaB 3 H 4 (400 Ci) in 100 l of 0.05 N NaOH at 30°C for 4 h, and the resulting radioactive oligosaccharide was purified as described previously (21). ␣ 1 AGP, RNase B, porcine thyroglobulin, and fetuin were also subjected to hydrazinolysis followed by N-acetylation according to the procedures described above.
Detection of Sulfate Ion Using Anion Chromatography-One nmol of oligosaccharides in fraction AR was heated in 6 N HCl at 110°C for 24 h followed by evaporation with methanol to remove HCl. The residue was dissolved in 100 l of water, and the two-fifths of the solution was applied to HPLC TSKgel IC-Anion-PW XL (4.6 mm inner diameter ϫ 3.5 cm, Tosoh Corp., Tokyo) to detect sulfate ion. Elution was performed with 1.5 mM potassium gluconate, 1.5 mM sodium borate, 5.8 mM boric acid, 4% acetonitrile, 3% 1-butanol, and 0.5% glycerol. The column temperature was maintained at 35°C at a flow rate of 1 ml/min.
Tetanus Toxoid-induced T Cell Proliferation-Tetanus toxoid-induced proliferation of normal human peripheral blood mononuclear cells was performed by using slight modifications of the previously described methods (22). Briefly, 10 ml of normal heparinized human blood was diluted with an equal volume of RPMI 1640 medium, and placed above 20 ml of Ficoll-Hypaque solution, and centrifuged at 800 ϫ g for 15 min at room temperature. Monocyte-enriched population located between upper and lower layers was carefully collected. The cells were washed twice with RPMI 1640 medium containing 10% autologous plasma by alternate suspension and centrifugation. The viable cells (2 ϫ 10 5 ) in 10% autologous plasma were incubated in 96-well microtiter plates with tetanus toxoid and various amounts of each oligosaccharide in a final volume of 200 l. After incubation for 6 days, the cultures were pulsed with 0.5 Ci of [ 3

H]thymidine for 6 h, and the amount of [ 3 H]thymidine incorporated into the cells was determined.
Measurement of the Amount of Cytokine in the Culture Medium of T Cells Proliferated by Tetanus Toxoid-The amount of IL-1␤ in the culture medium was measured at day 3 using an enzyme-linked immunosorbent assay kit for IL-1␤ (Otsuka Bioassay Research, Tokushima, Japan). The amount of IL-2 in the culture medium was measured at day 4 by counting the [ 3 H]thymidine incorporated into CTLL cells, whose growth is dependent on IL-2. Briefly, 5 ϫ 10 3 cells were incubated with 100 l of the culture medium of RPMI 1640 containing 10% FCS at 37°C for 18 h. The cultures were then pulsed with 0.5 Ci of [ 3 H]thymidine overnight, and the amount of [ 3 H]thymidine incorporated was determined as described above. For the quantitation, known amounts of rIL-2 were added to aliquots of the above assay mixture as standards.
Growth Inhibition of IL-1␤ Responsive T Cell Clone, D10-G4 by Oligosaccharides-In each well of a 96-well plate, 2 ϫ 10 5 of D10-G4 cells/50 l of RPMI 1640 containing 10% FCS were incubated with 50 l of 8 g/ml Con A (final concentration of 2 g/ml) and 50 l of 4 units/ml IL-1␤ (final concentration of 1 unit/ml). To this mixture, various concentrations of oligosaccharides were added and incubated at 37°C for another 2 days. The cultures were then pulsed with 0.5 Ci of [ 3 H]thymidine overnight, and the amount of [ 3 H]thymidine incorporated was determined as described above.
Binding of IL-1␤ to Oligosaccharides Immobilized on an NH 2 HPTLC Plate-Binding assays of IL-1␤ to oligosaccharides were performed using a slight modification of the previously described methods (23). An amino-bonded high performance silica gel plate (HPTLC Fertigplatten NH 2 F 254 ) was soaked in NaBH 3 CN solution and air-dried. Each oligosaccharide (1 nmol) was spotted onto the plate and left at room temperature (20 -25°C) for an appropriate length of time. The plate was washed three times with distilled water for 1 min and then soaked in saturated sodium bicarbonate solution containing 5% acetic anhydride to block the remaining active amino groups on the plate by N-acetylation. The plate was then overlaid with 5% bovine serum albumin (BSA) in 10 mM Tris-HCl, pH 7.8, containing 0.15 M NaCl and incubated at room temperature for 4 h. After being washed with PBS, the plate was overlaid with IL-1␤ solution in 1% BSA/PBS and incubated at 4°C for 16 h. After being washed five times with PBS, the plate was overlaid with rabbit anti-human IL-1␤ antibodies and incubated at 4°C for a further 2 h. The plate was then washed with PBS, overlaid with 125 Ilabeled goat anti-rabbit immunoglobulin (GϩM) antibodies, and incubated at room temperature for 2 h. The plate was then washed with PBS, dried, and exposed to XAR-5 x-ray film (Eastman Kodak Co.) in the dark at Ϫ80°C for 2-5 days.
Dot Blot Analysis-One or 10 g of each glycoprotein (THGP and bovine TSH) was spotted onto a nitrocellulose filter and air-dried for several hours. The filter was washed three times with PBS, soaked in 5% BSA/PBS, and incubated at room temperature for 4 h. After being washed with PBS, the filter was soaked in IL-1␤ solution in 1% BSA/ PBS and incubated at 4°C for 16 h. After being washed three times with PBS, the filter was soaked in the solution containing rabbit antihuman IL-1␤ antibodies and incubated at 4°C for 2 h. After a further wash with PBS, the filter was soaked in a solution containing biotinylated goat anti-rabbit immunoglobulins (GϩM) antibodies and incubated at room temperature for 2 h. The filter was then washed with PBS, incubated with ABC reagent at room temperature for 30 min, and washed with PBS again. The filter was incubated with a substrate solution (10 mM Tris-HCl buffer, pH 7.2, 0.3% 4-chloro-1-naphthol in methanol, 30% H 2 O 2 aqueous (5:1:0.01, v/v)) at room temperature for an appropriate length of time.
Binding of IL-1␤ of Glycosphingolipids on an HPTLC Plate-SM2, GM2, and sulfatide (1 nmol each) were spotted onto a silica polygram Sil G plate (Nagel, Germany). In order to prevent nonspecific binding, the plate was treated with 1% polyvinylpyrrolidone for 1 min and then soaked in PBS containing 1% BSA for 2 h at room temperature. The plate was incubated with IL-1␤ solution (0.1 g/ml) at 4°C for 18 h with gentle shaking. After washing three times with PBS, the plate was then incubated with rabbit anti-human IL-1␤ antibody solution at 4°C for 2 h. After washing three times with PBS, the plate was incubated with biotinylated goat anti-rabbit Ig(GϩM) antibody at room temperature for 1 h. Bound biotinylated antibody was detected as described above.

N-Linked Sugar Chains Obtained from THGP Inhibited Tetanus Toxoid-induced T Cell Proliferation-
The oligosaccharide fractions obtained from THGP by hydrazinolysis followed by N-acetylation were added to the culture medium of tetanus toxoid-specific T cell proliferation. The effect of these oligosaccharides on the proliferation reaction is shown in Fig. 1. The concentration required for 50% inhibition of tetanus toxoidinduced proliferation was around 10 M. At this concentration, neither the oligosaccharides liberated from ␣ 1 AGP nor those from RNase B inhibited tetanus toxoid-induced T cell proliferation (data not shown). In order to characterize the oligosaccharide showing the inhibitory activity, the oligosaccharide mixture obtained from THGP was fractionated. To facilitate detection of the oligosaccharides in further fractionation procedures, a small amount of tritium-labeled oligosaccharide mixture from THGP was added.
The oligosaccharide mixture was subjected to anion-exchange column chromatography with a Mono-Q HR5/5 column. As shown in Fig. 2A, oligosaccharides were separated into a neutral fraction (N) that eluted with 5 mM sodium acetate, pH 4.0, and an acidic fraction (A) that eluted with 500 mM sodium acetate, pH 4.0. When fraction A was exhaustively incubated with A. ureafaciens sialidase, 67% of it was converted to neutral components (named fraction AN), and the remainder (named fraction AR) was resistant to this enzymatic treatment (Fig. 2B). Approximately 80% of oligosaccharides in fraction AR were converted to neutral components (named fraction ARN) by mild methanolysis (Fig. 2C). These results suggest that the acidic nature of the majority of the oligosaccharides in fraction AR may be due to the presence of sulfate residues as it was previously reported that sulfate residues are removed from sugar moieties under these conditions (19). In order to confirm the presence of sulfate residues in fraction AR, this fraction was treated with 6 N HCl at 110°C for 24 h followed by anion chromatography using a TSK gel IC-Anion-PW XL column. The sulfate ion in the above reaction mixture was detected as shown in Fig. 3. This result indicates that the acidic nature of sialidase-resistant oligosaccharides is due to the sulfate residue. Based on the analytical data, about 3 nmol of sulfate ion was released from 1 nmol of oligosaccharide mixture in fraction AR under these experimental conditions. Remaining acidic oligosaccharides in fraction ARR in Fig. 2C were converted to the neutral components by heating in 0.1 N CF 3 COOH at 80°C for 2 h. However, the acidic nature of oligosaccharides in fraction ARR could not be determined due to the limited amounts of sample. Based on the radioactivities, the molar ratio of oligosaccharides in each fraction was calculated as follows: fraction N (10%), fraction A (90%), fraction AN (60%), fraction AR (30%), fraction ARN (25%), and fraction ARR (5%).
On the other hand, when fraction A was directly subjected to mild methanolysis in order to determine the ratio of oligosaccharides carrying only the sulfate residue, a very small part of the fraction was converted to a neutral oligosaccharide mixture (named fraction AMN), and most of the oligosaccharides remained acidic as indicated by the fraction AMR peak in Fig. 2D. The molar ratio of the oligosaccharides in the two fractions was calculated on the basis of their radioactivities, fraction AMN (3%) and fraction AMR (87%). These results indicate that 60% of the THGP oligosaccharides contain only sialic acid residues,

FIG. 2. Fractionation of oligosaccharides from THGP by an anion-exchange column.
Anion-exchange column chromatography was carried out using a fast protein liquid chromatography apparatus as described under "Materials and Methods." In order to facilitate detection of the oligosaccharides, a trace tritium-labeled oligosaccharide mixture (5 ϫ 10 6 cpm) from THGP was added. A, oligosaccharides liberated from THGP by hydrazinolysis; B, fraction A in A digested exhaustively with A. ureafaciens sialidase; C, fraction AR in B subjected to mild methanolysis treatment; D, fraction A in A subjected to mild methanolysis treatment. The arrows indicate the positions where the elution buffer was switched to 500 mM sodium acetate. 3% contain only sulfate residues, and 27% contain both sialic acid and sulfate residues as their acidic components.
Inhibition Activities of the Oligosaccharide Fractions-The inhibitory activity of each oligosaccharide fraction thus obtained on the T cell proliferation was examined, and the results are summarized in Fig. 4. Oligosaccharides in fraction A and fraction AR showed strong inhibitory activity. On the other hand, oligosaccharides in fraction ARN, which was converted to a neutral fraction by mild methanolysis treatment as described above, showed no inhibitory activity. These results indicate that the sulfate residue of oligosaccharides in fraction AR was essential for the expression of inhibitory activities. This interpretation is supported by the fact that fractions ARR, AMN, and AMR showed no inhibitory activity (Fig. 4). It is noteworthy that fraction AR showed only slightly more potent inhibitory activity than fraction A, whereas fraction A was separated into bioactive fraction AR and inactive fraction AN after sialidase treatment as described above. The reason for this has not yet been determined, although a possible explanation is that the presence of the sialic acid residues, which is not essential for proliferation-inhibitory activity, on the sugar chains in fraction A may enhance the inhibitory activity of the sulfated oligosaccharides. As compared with other inactive fractions, the neutral fraction N, which was reported to contain only high mannose-type oligosaccharides (24,25), had a slight inhibitory effect on the T cell proliferation at a concentration of ϳ100 M.
Inhibitory Mechanism of Oligosaccharides on T Cell Proliferation-To clarify the mechanisms of inhibitory activity of oligosaccharides in fraction AR, we changed the time of oligosaccharide addition. When added 24 h or more after the addition of tetanus toxoid, oligosaccharides in the fraction AR showed no inhibitory activity at all (Fig. 5). This result suggests that sulfated oligosaccharides in fraction AR acted at an early stage of antigen-specific T cell proliferation.
Therefore, we quantified the amount of IL-1␤ and IL-2 molecules in the culture medium of tetanus toxoid-activated T cells under the effect of various amounts of fraction AR, since these soluble mediators are known to play important roles at the early stage of antigen-specific T cell proliferation. As shown in Fig. 6, the amount of IL-1␤ in the T cell culture medium increased in parallel with the amount of fraction AR added. In contrast, the amount of IL-2 decreased on increasing the con-centration of oligosaccharides. A possible explanation for these interesting results is that fraction AR inhibited the binding of IL-1␤ to its receptor on the surface of T cells. As a result, the decline of signal transduction induced by IL-1␤ in T cells would reduce IL-2 production. This inhibition can be induced by bind- ing of the oligosaccharides to either IL-1␤ molecules or IL-1 receptors.
In order to find out whether the oligosaccharides in fraction AR can directly act on the IL-1␤ molecule or not, we examined the effect of fraction AR on the culture of an IL-1-responsive D10-G4 cell line (17,18). As shown in Fig. 7, both total oligosaccharides liberated from THGP (W in the figure) and fraction AR inhibited the proliferation of D10-G4 cells. However, oligosaccharides derived from other glycoproteins, ␣ 1 AGP and RNase B, did not show any effect at all (Fig. 7). These results suggest that the inhibition of antigen-specific T cell proliferation might be due to interaction of the unique oligosaccharides of THGP with IL-1␤.
Binding of IL-1␤ to Immobilized Oligosaccharides on an NH 2

FIG. 6.
Determination of the amounts of IL-1␤ and IL-2 in the T cell culture medium, which were activated by tetanus toxoid. E, amounts of IL-1␤ in the culture medium were determined using an enzyme-linked immunosorbent assay kit. q, amounts of IL-2 in the culture medium were measured using CTLL cells. The detailed procedures are described under "Materials and Methods." Inhibition of tetanus toxoid-induced T cell proliferation by fraction AR was performed as described in Fig. 4  4 ) were incubated with 2 g/ml Con A and 1 unit/ml human rIL-1␤ at 37°C. Each oligosaccharide fraction obtained from THGP (W, total oligosaccharides mixture; AR, oligosaccharides mixture in fraction AR), from ␣ 1 AGP, or from RNase B by hydrazinolysis followed by N-acetylation was added at the beginning of the incubation. Incubation was continued at 37°C for another 48 h, and then [ 3 H]thymidine incorporation was determined. Values are the means of the data obtained from three separate experiments. The bars indicate the standard deviations.
HPTLC Plate-We then examined whether IL-1␤ binds directly to the oligosaccharides of THGP by a thin layer overlay method as described previously (23). As shown in Fig. 8A, IL-1␤ bound to oligosaccharides in fractions A and AR but not to those in fractions AN, ARN, ARR, AMR, and AMN. These results are consistent with the data that only fractions AR and A could inhibit T cell proliferation (see Fig. 4). IL-1␤ also did not bind to the oligosaccharide fractions from thyroglobulin and ␣ 1 AGP, which did not inhibit the T cell proliferation (Fig.  8A). The radioactivities retained by immobilized oligosaccharides in Fig. 8A were quantified densitometrically, and the data are shown in Fig. 8B. The data clearly indicated that IL-1␤ bound only to the oligosaccharides in fractions A and AR de-rived from THGP. It must be stressed here that IL-1␤ very weakly bound to the porcine thyroglobulin oligosaccharides, which included sulfated biantennary sugar chains (26). The results indicate that oligosaccharides with not only sulfate residues but also additional structural features are recognized by the IL-1␤ molecule.
In order to elucidate the structure of the IL-1␤ ligand, fraction AR was applied to a W. floribunda agglutinin-agarose column, which is known to recognize a peripheral ␤-linked N-acetylgalactosamine residue (27), and separated into an unbound fraction (fraction I in Fig. 9A) and a bound fraction subsequently eluted with 100 mM N-acetylgalactosamine (fraction II in Fig. 9A). The percent molar ratio of these two frac-  Fig. 2; thyro, the oligosaccharide fraction obtained from porcine thyroglobulin; ␣ 1 AGP, that from ␣ 1 -acidic glycoprotein. B, quantification of the binding of IL-1␤ to oligosaccharide fractions immobilized on an NH 2 HPTLC plate. The amounts of IL-1␤ bound to the oligosaccharide fractions on the plate were determined by densitometry. Detailed procedures are described under "Materials and Methods." FIG. 9. Chromatography of oligosaccharides in fraction AR on a W. floribunda agglutinin-agarose column. A, fraction AR was subjected to a W. floribunda agglutinin-agarose column as described under "Materials and Methods." Arrows indicate the positions where the elution buffer was switched to that containing 100 mM N-acetylgalactosamine; B, rechromatography of oligosaccharides in fraction II in A after digestion with jack bean ␤-N-acetylhexosaminidase; C, binding of IL-1␤ to the oligosaccharides separated by the W. floribunda agglutinin column chromatography. Fractions I and II in A, fraction III in B, and fraction AR were immobilized on an NH 2 HPTLC plate. One nanomole of each oligosaccharide was spotted onto the plate. Detailed procedures for staining are described under "Materials and Methods." tions was 53 and 47%, respectively. After removing salt and N-acetylgalactosamine by passing through a Bio-Gel P-2 column, both the inhibitory activity on T cell proliferation and the binding ability to IL-1␤ were examined for the two fractions. As shown in Fig. 10, A and B, both activities resided exclusively in fraction II, and fraction II showed more potent inhibitory activity than fraction AR. Furthermore, removal of terminal Nacetylgalactosamine residues from fraction II by jack bean ␤-N-acetylhexosaminidase digestion (fraction III in Fig. 9B) abolished the binding ability to IL-1␤ completely (Fig. 9C). These data indicate that at least two elements, a sulfate residue and a ␤-N-acetylgalactosamine residue at the nonreducing termini of the oligosaccharides, are necessary for their binding to IL-1␤.
Binding of Glycoproteins and Glycosphingolipids with IL-1␤-In an attempt to clarify the relationship between the sulfate and N-acetylgalactosamine residues, we examined the binding of IL-1␤ to TSH because the SO 4 -4GalNAc group was present in the sugar chains of TSH (28). When we studied the interaction of THGP and TSH with IL-1␤, only THGP could be stained with IL-1␤ under the experimental conditions used (Fig. 11A). This suggests that IL-1␤ could not bind the SO 4 -4GalNAc group. Next, we examined the reactivity of IL-1␤ with various glycolipids containing sulfate and ␤-N-acetylgalactosamine residues. As shown in Fig. 11B, IL-1␤ did not bind to sulfatide at all. This is consistent with the results that the oligosaccharides of thyroglobulin did not react with IL-1␤ as shown in Fig. 8, because sulfatide contains only the SO 4 -3Gal group similar to thyroglobulin (26,29). On the other hand, it is quite interesting that IL-1␤ can bind weakly to SM2 but not to GM2 at all, although SM2 and GM2 have similar peripheral structures: the GalNAc␤134(R-3)Gal␤13, in which R is SO 4 for SM2 and is sialic acid for GM2 (30). Therefore, the GalNAc␤134(SO 4 -3)Gal␤13 group could be at least a part of the ligand for IL-1␤ molecule. DISCUSSION Muchmore's group (25,31) reported that the glycopeptides containing high mannose type sugar chains derived from THGP inhibited antigen-specific T cell proliferation by 50% in the concentration range 0.2-2 M. Dall'Olio et al. (7) reported that glycopeptides containing the complex type or high man-nose type sugar chains obtained by Pronase digestion of THGP inhibited the lymphocyte proliferation induced by mixed lymphocyte reaction. Conflicting evidence is that the glycopeptides obtained from ovalbumin by Pronase digestion, which should contain a series of high mannose type sugar chains together with a series of hybrid type sugar chains (2), did not show any inhibitory activity (7). Our data demonstrate, however, that oligosaccharides in fraction N, which contains high mannose type sugar chains (24,25), showed very little inhibitory activity even at the concentration of 100 M (Fig. 4). In contrast, oligosaccharides in fraction AR inhibited T cell proliferation by 50% at the concentration of 2 M.
We propose in this report that the immunosuppressive properties of THGP are expressed by the interaction of its carbohydrate moieties with IL-1␤. Our results demonstrate that oligosaccharides in fraction AR inhibited the proliferation of D10-G4 cells, which shows IL-1-dependent growth (17,18), and that IL-1␤ interacted specifically with the oligosaccharides in  10. Coexistence of the inhibitory activity and IL-1␤ binding activity in fraction AR and fraction II obtained by W. floribunda agglutinin column chromatography. A, effect of addition of the oligosaccharides fractionated by the W. floribunda agglutinin-agarose column on tetanus toxoid-induced lymphocyte proliferation. Fractions I and II are the same as in Fig. 9A. B, binding of IL-1␤ to each oligosaccharide fraction in A. The amount of IL-1␤ bound to each oligosaccharide fraction was determined by densitometry of the NH 2 HPTLC plate as described in Fig. 8. Detailed procedures are described under "Materials and Methods." fraction AR immobilized on the NH 2 HPTLC plate. By further fractionation of the oligosaccharides in fraction AR, it was found that the oligosaccharides containing ␤-N-acetylgalactosamine and sulfate residues specifically bind to IL-1␤. Two conflicting data were reported relating to our findings. Muchmore and Decker (13) reported that the N-linked sugar chains of uromodulin interacted with IL-1, although Moonen et al. (14) reported that THGP did not interact with soluble native cytokines and could only bind to denatured cytokines at low pH. Additionally, Fukushima et al. (32) demonstrated that IL-1␤ interacted with the glycosylphosphatidylinositol anchor. We recently found that IL-1␤ molecules radiolabeled with 125 I by Bolton-Hunter reagent, which reacted with N-terminal amino acid and lysine residues of the peptide portion (33), could no longer bind to THGP oligosaccharides. 2 This result suggests that N-terminal amino acid and/or lysine residues of IL-1␤ may play a role as the carbohydrate-binding site. In contrast, it was reported that even after the 125 I-labeling IL-1␤ molecules can bind to their receptors on the cell surface (34). Therefore, care must be taken when using an 125 I-labeled IL-1␤ as a probe to investigate the lectin-like activity of IL-1␤. The possibility that the inhibitory oligosaccharides bind to tetanus toxoid and consequently inhibit the T cell proliferation will be ruled out because toxoid-independent T cell proliferation induced by combination between phytohemagglutinin-L 4 and IL-1␤ was also inhibited by oligosaccharides in fraction AR (data not shown).
THGP has been found to contain more than 150 N-linked sugar chains, and the structures of only 30 oligosaccharides have been determined (35,36). Among them, two types of terminal sulfated elements were found, 4-O-sulfated GalNAc and 3-O-sulfated Gal (36). None of these are considered to be the actual ligand of IL-1␤ from our current study. These results indicate that not only the presence of a sulfate residue and a ␤-N-acetylgalactosamine residue but also a specific linkage(s) between the two groups is required for the ligand of IL-1␤. Based on the reactivity with glycosphingolipids, the GalNAc␤134(SO 4 Ϫ -3)Gal␤13 group could be considered as a part of an epitope for binding to the to IL-1␤ molecule. Therefore, structural determination of the remaining oligosaccharides of THGP, especially sulfated and ␤-N-acetylgalactosamine-containing N-linked sugar chains, will be required in order to elucidate the inhibitory mechanism of the IL-1␤ molecule and the physiological roles of the lectin-like property.