zyxwvut zyxwvutsrqpo Eur. J. Biochem. 203, 135-141 (1992) 0FEBS 1992 Glycosylation of interleukin-6 purified from normal human blood mononuclear cells zyxwvutsrq Raj B. PAREKH', Raymond A. DWEK', Thomas W. RADEMACHER', Ghislain OPDENAKKER' and Jo VAN DAMME' Glycobiology Unit, Department of Biochemistry, Oxford, England Rega Institute, Leuven, Belgium (Received July 11, 1991) - EJB 91 0913 Interleukin 6 (IL-6) is a glycosylated cytokine which is important in exerting cell-specific growthinducing, growth-inhibiting and differentiation-inducing effects. IL-6 produced in mammalian cell lines is heterogeneous, reflecting specific cell-type-dependent post-translational modifications. Native IL-6 was purified from human blood mononuclear cells and the oligosaccharides released, radiolabelled and sequenced by a combination of sequential exoglycosidase digestion using Bio-Gel P-4 high-resolution gel chromatography and acetolysis. N- and 0-linked glycans were found. The Nlinked glycans were sialylated di- and tri-antennary complex-type and oligomannose-type structures. However, the most predominant N-linked oligosaccharide was a small tetrasaccharide with the sequence Mana6Manp4GlcNAcp4GlcNAc. This is the first report of this structure on a circulating glycoprotein. This structure has only previously been reported to be present on the syncytiotrophoblast of human placenta. The presence of the oligomannose structures and the mannose-terminating tetrasaccharide on IL-6 may be important in maintaining a high local concentration of the cytokine while limiting its systemic serum level via interaction with soluble mannose-binding serum lectins. Interleukin-6 (IL-6) is a cytokine with multiple biological functions including B-cell stimulation for the production of antibodies, growth-promoting activity for plasmacytoma and hybridoma cells, induction of acute phase responses and Tcell activation (for reviews see Le and Vilcek, 1989; Van Snick, 1990). IL-6 is the translation product of a single-copy gene located on the long arm of human chromosome 7 (Bowcock et al., 1988); the cDNA of IL-6 has been cloned in several independent laboratories (Haegeman et al., 1986; Zilberstein et al., 1986; Hirano et al., 1986). IL-6 gene expression is dependent on both the inducer substance and the cell type involved. The most potent physiological inducer discovered so far is IL-1 (Van Damme et al., 1987b). Other cytokines, e. g. tumor necrosis factor as well as double-stranded RNA, viral and bacterial products such as endotoxin are also able to induce significant levels of this cytokine (Van Damme et al., 1987b, 1989). Glucocorticoids, indirectly controlled by IL6-induced adrenocorticotrophin hormone release, are potent inhibitors of 1L-6 expression (Kohase et al., 1987). Regulated induction of lL-6 is monitored by factors that also control the transcription of tumor necrosis factor, IL-8 and granulocyte colony stimulating factor (Akira et al., 1990; Shimizu et al., 1990; Zhang et al., 1990; Libermann and Baltimore, 1990). The cellular responses to 1L-6 have been intensely investigated. IL-6 interacts with a cellular receptor, belonging to the immunoglobulin superfamily (Yamasaki et al., 1988); this interaction results in the activation of a 130-kDa molecule zyxwvu that transmits the signal to the intracellular compartment (Taga et al., 1989). This results in the activation of (a) specific transcription factor(s) that interact(s) with IL-6-specific promoter/enhancer elements of the responsive genes (e. g. of the acute-phase reactants) (Majello et al., 1990; Li et al., 1990; Ito et al., 1989; Oliviero and Cortese, 1989). Although the primary structure of the translation product of the IL-6 gene is known, little is known of its post-translational processing. IL-6 derived from human fibroblasts can be processed by phosphorylation at serine residues (May et al., 1988b). Further, two potential N-glycosylation sites are located at amino acid residues 46 and 145 of the most abundant natural form from human leukocytes (Haegeman et al., 1986; Zilberstein et al., 1986; Hirano et al., 1986; Van Damme et al., 1988). No carbohydrate structures have so far been described, probably due to the scarcity of pure natural product. In this paper the structures from the predominant forms of monocytic IL-6 are defined. Carbohydrate sequencing was successfully performed at the picomolar level and 98% of the occurring structures present could be determined. MATERIALS AND METHODS Production and purification of human IL-6 from peripheral blood mononuclear cells was as described elsewhere (Van Damme et al., 1987a). Briefly, buffy coats from pooled blood donations were treated with hydroxyethyl starch (Plasmasteril, Fresenius AG, Bad Homburg, FRG) to remove erythrocytes. The mononuclear cells were isolated by gradient centrifugation on Ficoll/sodium metrizoate (Lymphoprep, Nyegaard, Oslo, Norway), suspended in RPM1/2% fetal calf serum and stimulated with 10 pg/ml concanavalin A, After 48 h, the cell culture media were harvested and the IL-6 purified. Essentially a five-step purification schedule was used : zyxwvutsrqpo Correspondence to T. W. Rademacher, Glycobiology Unit, Department of Biochemistry, South Parks Road, Oxford, OX1 3QU, England Abbreviations. IL-6, interleukin 6 ; IL-8, interleukin 8. Enzymes. u-Mannosidase (EC 3.2.1.24); P-galactosidase (EC 3.2.1.23); p-N-acetylhexosaminidase (EC 3.2.1.52); sialidase (EC 3.2.1.18). 136 zyxwvut zyxwvuts zyxwvutsrqpo concentration and batch adsorption on silicic acid (Mallickrodt Inc. Paris, KY), followed by elution with ethylene glycol (50% in 1.4 M NaC1) and 0.3 M glycine pH 2; antibody affinity chromatography using a goat polyclonal antiserum coupled to CNBr-activated Sepharose 4B and acid elution (pH 2); gel filtration chromatography on Ultrogel AcA 54 eluting at about 30 kDa (Pharmacia/LKB, Bromma, Sweden); cationexchange chromatography on Mono S (Pharmacia, Uppsala, Sweden) in 50 mM formate pH 4.0 and elution at 0.45 M NaCl; and finally reverse-phase HPLC on a C1 25-nm pore size column eluting at 38% acetonitrile. In some experiments IL-6 was purified by adsorption to silicic acid and by monoclonal antibody affinity chromatography. Structural analysis of the NH2-terminal sequences of native human leukocyte 1L-6 Homogeneity of the purified material was verified by (a) SDSjPAGE (linear 8 - 20% gradient gel and silver staining); (b) by determination of the specific biological activity in the hybridoma growth assay and (c) NH2-terminal amino acid sequence analysis on a gas-phase sequencing apparatus (model 477A, Applied Biosystems, Foster City, CA, USA) equipped with an on-line phenylthiohydantoin derivative analyser (model 120 A) (Van Damme et al., 1987a). zyxwvuts Release of oligosaccharidesfrom IL-6 Native human IL-6 (z20 pg) was rendered salt-free by dialysis against double-distilled water. After lyophilisation, the oligosaccharides covalently attached to IL-6 were released, labelled and isolated as reported previously (Ashford et al., 1987). The radiolabelled oligosaccharides were fractionated using a combination of high-voltage paper electrophoresis and Bio-Gel P-4 ( z 400 mesh) gel filtration chromatography (Parekh et al., 1989a, b). Structural analysis of the radiolabelled, desialylated oligosaccharides obtained after gel filtration chromatography was performed using a combination of exoglycosidase digestion and controlled acetolysis, also as described previously (Parekh et al., 1989a, b; Ashford et al., 1987; Olafson et al., 1990). Insufficient quantity of oligosaccharide was available for methylation analysis. Standard oligosaccharides were prepared from glycoproteins by hydrazinolysis with subsequent reduction with NaB3H4, as described previously (Parekh et al., 1989a). For all oligosaccharide structures (Fig. 3), other than H, elution positions corresponded to those of the appropriate standard alditol (Parekh et al., 1989a). In addition the disaccharide GalB3GalNAcoT (where OT subscript refers to the reducing terminus) was prepared from bovine fetuin, and GalNAcoT was obtained by reduction of N-acetylgalactosamine. Identification of the radiolabelled reducing terminus of each oligosaccharide as either GlcNAcoT or GalNAcoT was based on comigration of the unknown monosaccharide alditol with standard GlcNAcOT or GalNAcoT during high-voltage paper electrophoresis in borate buffer, as described previously (Walsh et al., 1989). RESULTS Purification and analysis of human leukocyte IL-6 Peripheral blood mononuclear cells from pooled human buffy coats were stimulated for 48 h with 10 pg/ml of concanavalin A for the production of interleukins (Van Fig. 1. Staining analysis of natural monocytic 1L-6. Arrows indicate the predominant species from which the carbohydrate structures were derived. Molecular masses of standards are indicated on the left in kDa. Damme et al., 1988). IL-6 was purified as described in Materials and Methods and the resulting glycoproteins analysed. Fig. 1 shows the result of the silver-stained product: the IL-6 appears as two predominant glycoprotein bands of 24 kDa and 28 kDa. Titration of the purified product yielded lo9 U/ mg material, a value compatible with the specific activities of IL-6 purified from fibroblasts (Van Damme et al., 1987a). From SDSjPAGE analysis, we could deduce that the 28-kDa and 24-kDa bands were present in almost equimolar amount. By automated Edman degradation the native Ala-Pro-ValPro-Pro-Gly-Glu-Asp sequence and NH2-truncated forms (lacking one or two amino acid residues) of IL-6 were found. The relative percentage occurrence of the intact form was 70%, and 7% and 23% for the NH2-truncated forms lacking one and two residues, respectively. This composition was confirmed with leukocyte-derived IL-6 obtained after purification to homogeneity by monoclonal antibody affinity chromatography. Monosaccharide sequence analysis For structural analysis of the oligosaccharides released from native human IL-6, an aliquot of the total mixture of reduced, radiolabelled oligosaccharides isolated from IL-6 was first subjected to high-voltage paper electrophoresis. The resulting profile is shown in Fig. 2a. The relative amount of neutral and acidic components was determined by the recovery of radioactivity from the paper (Table 1). Treatment prior to electrophoresis of an aliquot of the mixture of oligosaccharides with the sialidase from Arthrobacter ureafaciens caused essentially complete conversion of the acidic oligosaccharides to neutral ones (Fig. 2b). This indicates that sialic acid in linkage a2 3 or a2 -+ 6 is responsible for the acidic nature of the IL-6-derived oligosaccharides. The small amount of acidic radioactivity remaining after sialidase is commonly found and zyxw --f zyxwvutsrqponmlkj zyxwvutsr z zyxwv 137 1 3 + SL (a1 0 Table 1. Relative amounts of different types of oligosaccharidesreleased from human IL-6. The fractions refer to the peaks in Fig. 3. The last column gives their positions of elution from the P-4 column relative to standard glucose oligomers, i.e. 16-17 indicates that the fraction is eluted between (Glc),, and ( G ~ C ) ~numbers ,; in parentheses refer to the peak positions. Peaks E, F, G and H are the oligomannose type; peaks A, B, C and D contain complex-type oligosaccharides; peak E probably contains (Man)8(GlcNAc)2since it is sensitive to a l 2-mannosidase; peak I is an 0-linked structure. Values at the bottom for neutral and acidic oligosaccharides were obtained following highvoltage electrophoresis. Fraction Occurrence zyxwvutsrq zyxwvutsrqponmlkji cf. total x + ._ .-> + " B ._ TI IY P-4 position N-linked 0-linked % 1 5 12 8 1 53 2 11 26 17 2 4 4 32 - 100 22 78 46 53 100 2 2 15 Neutral Acidic 16-17 14-15 (14.5) 13-14 (13.5) 12-13 (12.5) 11- 12 (1 1.5) 9-10( 9.8) 8- 9 ( 8.9) 6- 7 ( 6.5) 3- 4( 3.5) I zyxwvutsr zyxwvutsrqp enzyme d removed one residue. Negative reactions which are performed on an adhoc basis to confirm certain structures are not reported in detail (e.g. structures C and D were non0 Distance 10 from origin 20 (cm] 30 reactive with a-fucosidase). Assignment of linkages and anomers is from the literature data on the exoglycosidases Fig. 2. High-voltage radioelectrophoretogramsof the oligosaccharides used. The arrangement of the mannose residues in structures recovered from native human IL-6 before (a) and after (b) exhaustive F and G are from the known biosynthetic pathway in mamdigestion with sialidase (ex. Avthvobacter ureafaciens). Neutral (N) and malian cells. A further discussion of other possible mannose acidic (A) oligosaccharides were recovered by elution with water. branching patterns is given elsewhere (Williams et al., 1991). L = lactitol; SL = 3'(6')-sialyl-lactitol. Identification of each reducing terminus monosaccharide (Walsh et al., 1989) was based on its co-migration with either GlcNAcoT or GalNAco, during high-voltage paper electrophoresis in borate buffer (data not shown). In the case can be due to either radiochemical blank or incomplete of fraction H, controlled acetolysis was used to establish the sialidase digestion when minute amounts (1 - 2 nmol in total) 6-linkage of the non-reducing-terminal a-linked mannose. of substrate are present. Under the acetolysis conditions used, cleavage of the Manal-6 Following sialidase treatment, the total pool of oligo- linkage is essentially complete, while cleavage of the Manal-3 saccharides (i.e. the neutral and desialylated acidic together) linkage is not significant (Parekh et al., 1989b). Almost all of were separated by Bio-Gel P-4 ( ~ 4 0 0mesh) gel filtration the oligosaccharide fraction H was converted by controlled chromatography (Fig. 3). Individual fractions were pooled acetolysis from a hydrodynamic volume corresponding to as indicated and the relative amounts are given in Table 1. (G1c)6.3(Fig. 5a) to a product with a hydrodynamic volume Fractions B, C, D, F, G, H, and I were further analysed by corresponding to ( G ~ c ) , . ~(Fig. 5 b), indicating that the sequential exoglycosidase digestion. Assignment of glycosyl acetolysis caused the removal of one non-reducing terminal residue sequence and the anomeric configuration of individual mannose residue. The product of hydrodynamic volume corglycosidic linkages is based on the reported specificities of responding to ( G ~ c ) , ,co-elutes ~ with authentic Manp4Glcthe individual exoglycosidases, the change in hydrodynamic NAcf14GlcNAcoT(Parekh et al., 1989a). volume induced by each exoglycosidase (Parekh et al., 1989a) and the hydrodynamic volumes of standard oligosaccharides. Each oligosaccharide fraction was sensitive to exoglycosidases DISCUSSION when these were used in the order indicated in Fig. 4. The The most abundant secreted form of fibroblast- and number of residues removed after each digestion can be obtained by counting the residues to the left of each dotted leukocyte-derived IL-6 is a glycoprotein containing 3 85 amino line. For example, in structure F, enzyme g removed one acids with two potential N-glycosylation sites located at resiresidue. Enzyme c acting on the product of enzyme g removed dues 46 and 145 respectively. Processing of the primary transfour residues. Enzyme d acting on the product of enzyme c lation product includes the cleavage of a signal peptide of 27 removed one residue, and enzyme e acting on the product of amino acids, phosphorylations at serine residues and N- and zy 138 zyxwvut zyxwvutsr zyx zyxw P 201918 17 16 15 1L 13 12 11 10 9 3 2 1 -1 zyxwvut zyxwvutsrqp Retention txie Fig. 3. Bio-Gel P-4 (zz 400 mesh) gel filtration chromatogram of the oligosaccharides (neutral and desialylated acidics) derived from native human IL-6. Individual fractions were pooled as indicated, the relative ratios determined (Table 1) and, in the cases of fractions B, C, D, F, G, H, and I, analysed further. Numbers at the top refer to the elution positions of co-applied unreduced (non-radioactive) standard glucose oligomers, i.e. 20 = elution position of ( G I c ) ~The ~ . hydrodynamic volume of the radiolabelled oligosaccharides is measured relative to these standard glucose oligoniers. 0-linked glycosylation as determined by incorporation of radiolabelled monosaccharides, tunicamycin inhibition and N- and 0-glycanase enzymatic digestion experiments (Santhanam et al., 1989). Furthermore, NH,-terminal clipping results in truncated IL-6 forms lacking one or two amino acid residues. This truncation is, however, not explanatory for the observed difference in molecular mass forms in denaturing SDSjPAGE analysis. This study concerns the characterisation of the carbohydrate structures in two predominant forms of native human monocytic IL-6. It should be emphasized that other minor ‘glycoforms’ were found to be present as secreted IL-6 species (data not shown). To date no compositional or structural data of the glycan structures have been published of either recombinant or natural IL-6. Table 1 shows that 46% of the N-linked glycans were neutral, 40% oligomannose, 53% were sialylated complextype oligosaccharides. In the total pool 22% of the structures were neutral and 19% oligomannose. Therefore neutral complex-type could at most make up 3% of the total sugar recovered. There was too little sample to provide an analysis of the neutral pool separately from the total asialo pool in order to confirm the presence of this small quantity of neutral complex glycan. The quantity of peak A was also insufficient for structural analysis; however, its P-4 position is consistent with a triantennary oligosaccharide. Fig. 4 shows several unusual features of the monosaccharide sequences from IL-6. First, the presence of the oligomannose structures (F and G) are uncharacteristic of a circulating glycoprotein (note that we analysed the secreted forms of cultured monocytes). Unless these residues are buried or inaccessible (those on IgM are localized to the tail fragment and are probably inaccessible; Wormald et al., 1991), interaction with various circulating lectins (e. g. mannan binding protein) or cellular receptors for mannosides (Rademacher et al., 1988 a) would be expected. The presence of oligomannose zyxwvu oligosaccharides may ensure a short circulating survival time for this potent cytokine which may be intended to work in a local immune network, rather than systemically. Alternatively, the presence of oligomannose structures on a subset of IL-6 molecules might target these to local action, whereas another subset might contribute to the systemic effects such as the induction of fever and the acute-phase reaction. Second, the finding of a linear tetrasaccharide was also unexpected (structure H, Fig. 4). We have previously found the same structure on the glycocalyx of the syncytiotrophoblast of human placenta (Arkwright et al., 1991). Interestingly, this tetrasaccharide does not bind to concanavalin A (Arkwright et al., 1991). Its affinity to other mannose binding lectins (i.e. mannan-binding protein) is at present unknown. Finally, the presence of a small amount of monogalactosylated structure is also unusual (structure D). To date IgG is unique in being the only known circulating glycoprotein to contain this residue (Rademacher et al., 1988a). On IgG these residues are buried and inaccessible as GlcNAc residues on glycoproteins are potent targets for anti-ClcNAc antibodies which are formed in the response to many infectious organisms such as group A streptococcus (Rademacher et al., 1988a; Rook et al., 1988). Exposed GlcNAc is however common on N-linked oligosaccharides of central-nervous-system glycoproteins (Wing et al., 1991, unpublished results) and nuclear proteins; in the latter case it is 0-linked to serine or threonine (Jackson and Tjian, 1988). IL-6 polypeptides are subject to extensive post-translational modifications. In general, multiple species of IL-6 can be resolved by SDSjPAGE under denaturing conditions: a triplet of molecular mass 23-25 kDa and another triplet of mass 28-30 kDa (Van Damme et al., 1988; May et al., 1989; and data not shown). Which species predominates, depends on the cellular source (i.e. fibroblast, monocytes, endothelial cells, plasma, synovial fluid have all been shown to have Fraction zyxwvuts zyxw zy zyxwv Structure A Gal0134 B I I I Gal0144 A A b A GlcNAcB1+2 I C Man0144 I I a Gal0134 I I I I I II GlcNAcOlh4 I I e A A A I t I Man8144 , I ? 1; C At- A A I zyxwvut GlcNAcOl&4 I I G l c N A c O l ~ 2 Mana 1 a 6 GlcNAc,, , d cl GalB1+4 139 f I I I C ;, : - - _ - - _ _ - I _ - . 3 Mana 1 ' %6 I I I I I Fuca 1 ' i zyxwvutsrqponmlkjihg zyxwvutsrqp A A G l c N A c O l ~ 2 Mana 1 ' <6 ' ' GIcNAc01~2 Mana l? 4 A GlcNAc,, I I I IAc01&4 I I a b \6 Mana lf 4 Mana 1 ' 3 %6 ' zy 4 4 I I I I 3 Man0144 I I Mana l? I I C Mana 1 \I2 G U 3 A .<- nrfl. Ma{.- I Mancx 1 ? C A H Mana 1 I I I e A A I GIcNAc, I I GlcNAcR1&4 d e A A 6 Man01+4 zyxw GlcNAcOl+4 d ? 6 Man0144 , a Mana 1 zyxw e C Mana 1 GIcNAc,, GlcNAc01+4 I I I C d e GlcNAc,, GIcNAc,, A 1 I Gal81fr 3 GalNAc,, il Fig. 4. Proposed structures of the neutral and desialylated oligosaccharides derived from native human 1L-6. Structural analysis was perforincd on individual oligosaccharide fractions (Fig. 3) by using sequential exoglycosidase digestion with the exoglycosidases indicated, followed by separation of the products using Bio-Gel P4 ( z400 mesh) gel filtration chromatography, this was followed by controlled acetolysis, as dcscribed in Materials and Methods. Changes in the hydrodynamic volume of the oligosaccharide fractions were effected by cxoglycosidases when used in the following order: oligosaccharide B, a-b-c-d-e-f; oligosaccharide C, a-b-c-d-e; oligosaccharide D, a-b-c-d-e; oligosaccharide F, g-c-d-e; oligosaccharide G, c-d-e; oligosaccharide H, c-d-e; oligosaccharide I, h. The enzymes used were: (a) jack bean P-galactosidase; (b) Streptococcus pneumoniae b-N-acetylhexosaminidase; (c) jack bean a-mannosidase; (d) Achatina fulica P-mannosidase; (e) jack bean /I-Nacetylhexosaminidase; (f)bovine epididymal a-fucosidase; (g) Aspergillusphoenicis al-2-specific mannosidase; (h) bovine testes P-galactosidase. The deduced points of hydrolysis of each structure by individual exoglycosidases are indicated. Oligosaccharides B, C, D, F, G, and 1 have the same hydrodynamic volume (to within 0.05 glucose unit) as the identical standard oligosaccharides. Each reducing-terminus monosaccharide (i.e. whether GIcNAc~Tor GalNAcoT) was identificd by its ability to co-migrate with either standard GlcNAc,, or GalNAc,, during high-voltage papcr electrophoresis in borate buffer (Walsh et al., 1989). 140 zyxwv zyxwvutsrqp zyxwvutsrq zyxwvutsrqpo zyxwvutsrqpo zyxwvutsrq 10987 6 5 L 4.1.14.1 4 .1 I 3 J. n 2 4 1 4 A’ I (Rook et al., 1991), as well as in Castleman’s disease, cardiac myxoma and mesangial proliferative glomerulonephritis all have the agalactosyl IgG glycoform type (unpublished data). These results suggest that IL-6 may affect either the glycosyltransferase levels of B cells or cause a unique subset of B cells with this phenotype to secrete at a higher rate. In support of the former, it has recently been shown that myeloma cells grown in the presence of IL-6 change their cellsurface glycosylation patterns (Nakao et al., 1990). The elevated serum levels of IL-6 in a number of disease states suggests either an increased rate of synthesis (constitutive) or a prolonged circulatory lifetime. It will be of interest in the future to determine if IL-6 expressed in a variety of cell lines still contains these unusual oligosaccharides and whether or not those glycoforms of IL-6 containing these monosaccharide sequences will constitute a distinct subset of IL-6 molecules with unique functions. In particular, nonglycosylated forms found in serum could act as ‘anticytokines’ (Rademacher et al., 1988a). Further, an analysis of the glycosylation of IL-6 from a number of disease states, in particular, IL-6 produced in the joints of rheumatoid arthritis patients, will be important in our understanding of the proposed dysregulated autocrine loop in that disease. zyxwvutsr zyxwvuts Retention time Fig. 5. Bio-Gel P4 (s400 mesh) gel filtration chromatogram of oligosaccharide fraction H before (a) and after (b) controlled acetolysis. Numbers at the top refer to the elution positions of co-applied unreduced (non-radioactive) standard glucose oligomers, i.e. 10 indicates the elution position of (G~C)~,,. different patterns), on cell culture conditions and on the purification methods (Van Damme et al., 1987a, 1988; May et al., 1988a). A 45-kDa IL-6 form is present in rheumatoid synovial fluid, produced by endothelial cells and a similar species is found circulating in human plasma (May et al., 1989). May and co-workers have proposed that all species of 1L6 contain 0-glycosylation but that N-glycosylation is variable (see May et al., 1991, and references therein). With two potential N-linked sites, four different species are possible. Further, splitting of these by differential phosphorylation and/or 0glycosylation may account for the final banding patterns (May et al., 1989). The data in Table 1 show that we isolated one 0-linked glycan/N-linked glycan. 0-Glycans are found to be released, albeit in variable yield, by the hydrazinolysis conditions of Ashford et al. (1987) as discussed previously (Walsh et al., 1989). The data is consistent with half of the IL-6 molecules containingjust an 0-linked glycan ( z23 - 25-kDa forms) and the other half containing two N-linked glycans and one 0linked glycan ( z 28 - 30-kDa fornis). Glycopeptide site analysis on individual purified IL-6 glycoforms will, however, be necessary to confirm this arrangement of glycans. Dysregulation of IL-6 is now thought to contribute to the pathogenesis of a number of disease states. In particular, multiple myeloma/plasmacytoma, Castleman’s disease, cardiac myxoma and mesangia1 proliferative glomerulonephritis are all associated with elevated IL-6 levels (Le and Vilcek 1989; Hirano et al., 1990; Van Snick, 1990). Transgenic IL-6 mice are characterized by kidney disease and plasmacytosis (Suematsu et al., 1989) and IL-6 is a growth factor for murine plasmacytomas (Van Damme et al., 1987b). Pristane injected into mice leads to high constitutive levels of IL-6 and contributes to myeloma formation (Nordan and Potter, 1986; Rook et al., 1991). Rheumatoid arthritis has also been proposed to be associated with IL-6 dysregulation (Le and Vilcek, 1989; Hirano et al., 1988, 1990; Van Snick, 1990; Houssiau et al., 1988). In this disease the IL-6 synthesis rate in the joint is increased as is the case for its physiological inducer IL-1 (Van Damme et al., 1987b). Interestingly, the IgG produced in rheumatoid arthritis (Rademacher et al., 1988a) and transgenic IL-6 mice JVD and GO are research associates of the Belgian Nationaal Fonds voor Wetensckappelijk Onderzoek. GO was a recipient of a research grant of the Flemish Community, Rezsbeurzenwedstr&f.The Glycobiology Unit is supported by Monsanto. REFERENCES Akira, S., Isshiki, H., Sugita, T., Tanabe, O., Kinoshita, S., Nishio, Y., Nakajima, T., Hirano, T. & Kishimoto, T. (1990) EMBO J . 9, 1897- 1906. Arkwright, P. D., Redman, C. W. G., Williams, P., Dwek, R. A. & Rademacher, T. W. (1991) Placenta, in the press. Ashford, D., Dwek, R. A,, Welply, J. K., Amatayakul, S., Homans, S. W., Lis, H., Taylor, G. N., Sharon, N. & Rademacher, T. W. (1987) Eur. J. Biochem. 166, 311 -320. Bowcock, A. M., Kidd, J. R., Lathrop, M., Danshvar, L., May, L. T., Ray, A., Sehgal, P. B., Kidd, K. K . & Cavalisforza, L. L. (1988) Genomics 3, 8-16. Haegeman, G., Content, J., Volckaert, G., Derynck, R., Tavernier, J. & Fiers, W. (1986) Eur. J. Biockein. 159, 625-632. Hirano, T., Yasukawa, K., Harada, H., Taga, T., Watanabe, Y., Matsuda, T., Kashiwamura, S., Nakajima, K., Koyama, K., Iwamatsu, A., Tsunasawa, S., Sakiyama, F., Matsui, H., Takahard, Y . ,Taniguichi, T. & Kishimoto, T. (1986) Nature 324, 73 - 76. Hirano, T., Matsuda, T., Turner, M., Miyasaka, N., Buchan, G., Tang, B., Sato, K., Shimizu, M., Maini, R., Feldmann, M. & Kishimoto, T. (1988) Eur. J. Imrnunol. 18, 1797-1801. Hirano, T., Akira, S., Taga, T. & Kishimoto, T. (1990) Immunol. Today 11, 443-449. Houssiau, F. A,, Devogelaer, J. P., Van Damme, J., Nagant de Deuxchaisnes, C. & Van Snick, J . (1988) Arthritis Rheum. 31, 784- 788. Ito, T., Tanahashi, H., Misumi, Y. & Sakaki, Y. (1989) Nucleic Acids Res. 17, 9425-9435. Jackson, S. P. & Tjian, R. (1988) Cell 55, 125- 133. Kohase, M., Henriksen-DiStefano, D., Sehgal, P. B. & Vilcek, J. (1987) J . Cell Physiol. 132, 271 -278. Le, J. & Vilcek, J. (1989) Lab. Invest. 61, 588-6602, Li, S.-P., Liu, T.-Y. & Goldman, N. D. (1990) J . Biol. Ckem. 265, 4136-4142. Libermann, T. A. & Baltimore, D. (1990) Mol. Cell. Biol. 10, 23272334. Majello, B., Arcone, R., Toniatti, C. & Ciliberto, G. (1990) EMBO J . 9, 457 - 465. zyxwvu zyxwvutsrq z zyxwvutsr zyxw zyxwvutsrqpon 141 May, L. T., Ghrayeb, J., Santhanam, U., Tatter, S. B., Sthoeger, Z., Helfgott, D. C., Chiorazzi, N., Grieninger, G. & Sehgal, P. B. (1988a) J . Biol. Chem. 264,7760-7766. May, L. T., Santhanam, U., Tatter, S. B., Bhardwaj, N., Ghrayeb, J. & Sehgal, P. B. (1988b) Biochem. Biophys. Res. Commun. 152, 1144 - 1150. May, L. T., Santhanam, U., Tatter, S. B., Ghrayeb, J. & Sehgal, P. B. (1989) Ann. N . Y. Acud. Sci. 557, 114-122. May, L. T., Santhanam, U. & Sehgal, P. B. (1991) J . Biol. Chem. 266, 9950- 9955. Nakao, H., Nishikawa, A,, Karosuno, T., Nishiura, T., Iida, M., Kanayama, Y., Yonezawa, T., Tarui, S. & Taniguchi, N. (1990) Biochem. Biophys. Res. Commun. 172, 1260- 1266. Nordan, R. P. &Potter, M. (1986) Science 233, 566-569. Olafson, R. W., Thomas, J. E., Ferguson, M. A. J., Dwek, R. A., Chaudhuri, M., Chang, K.-P. & Rademacher, T. W. (1990) J . Biol. Chem. 21, 12240-12247. Oliviero, S. & Cortese, R. (1989) EMBO J . 8, 1145-1151. Parekh, R. B., Dwek, R. A,, Thomas, J. R., Rademacher, T. W., Opdenakker, G., Wittwer, A. J., Howard, S. C., Nelson, R., Siegel, N. R., Jennings, M. G., Harakas, N. H. & Feder, J. (1989a) Biochemistry 28, 1644- 7662. Parekh, R. B., Dwek, R. A,, Rudd, P. M., Thomas, J. R., Rademacher, T. W., Warren, T., Wun, T.-C., Hebert, B., Reitz, B., Palmier, M., Ramabhadran, T. & Tiemeier, D. C. (1989b) Biochemistry 28, 1670- 7679. Rademacher, T. W., Parekh, R. B. & Dwek, R. A. (1988a) Annu. Rev. Biochrm. 57, 785 - 838. Rademacher, T. W., Parekh, R. B., Dwek, R. A., Isenberg, D., Rook, G., Axford, J. S. & Roitt, I. (1988b) Springer Sem. Zmmunopath. 10, 231 -249. Rook, G. A. W., Steele, J. & Rademacher, T. W. (1988) Ann. Rheum. Dis. 47, 247-250. Rook, G., Thompson, S., Buckley, M., Elson, C., Brealey, R., Lambert, C., Whyte, T. & Rademacher, T. W. (1991) Eur. J. Immunol. 21,1027-1032. Santhanam, U., Ghrayeb, J., Sehgal, P. & May, L. T. (1989) Arch. Biochern. Biophys. 274, 161 - 170. Shimizu, H., Mitomo, K., Watanabe, T., Okamoto, S. & Yamamoto, K.-I. (1990) Mol. Cell. Biol. 10, 561 -568. Suematsu, S., Matsuda, T., Aozasa, K., Akira, S., Nakano, N., Ohno, S., Miyazaki, J.-I., Yamamura, K.-I., Hirano, T. & Kishimoto, T. (1989) Proc. Nut1 Acad. Sci. U S A 56, 7547-7551. Taga, T., Hibi, M., Hirata, Y., Yamasaki, K., Yasukawa, K., Matsuda, T., Hirano, T. & Kishimoto, T. (1989) Cell%', 573581. Van Damme, J., Cayphas, S., Van Snick, J., Conings, R., Put, W., Lenaerts, J.-P., Simpson, R. J. & Billiau, A. (1987a) Eur. J. Biochem. 168, 543-550. Van Damme, J., Opdenakker, G., Simpson, R. J., Rubira, M. R., Cayphas, S., Vink, A,, Billiau, A. & Van Snick, J . (1987b) J . Exp. Med. 165,914-919. Van Damme, J., Van Beeumen, J., Decock, B., Van Snick, J., De Ley, M. & Billiau, A. (1988) J . Zmmunol. 140, 1534-1541. Van Damme, J., Schaafsma, M. R., Fibbe, W. E., Falkenburg, J. H. F., Opdenakker, G. & Billiau, A. (1989) Eur. J . lmmunol. 19,163 168. Van Snick, J. (1990) Annu. Rev. Immunol. 8, 253-278. Walsh, F. S., Parekh, R. B., Moore, S. E., Dickson, G., Barton, C. H., Gower, H. J., Dwek, R. A. & Rademacher, T. W. (1989) Development (Comb.) 105,803-811. Williams, P., Wormald, M. R., Dwek, R. A., Rademacher, T. W., Parker, G. F. & Roberts, D. R. (1991) Biochim. Biophys. Acta 1075,146-153. Wormald, M. R., Wooten, E. W., Bazzo, R., Edge, C. J., Feinstein, A,, Rademacher, T. W. & Dwek, R. A. (1991) Eur. 1.Biophys., in the press. Yamasaki, K., Taga, T., Hirata, Y., Yawata, H., Kawanishi, Y., Seed, B., Taniguchi, T., Hirano, T. & Kishimoto, T. (1988) Science 241, 825 - 828. Zhang, Y., Lin, J.-X. & Vilcek, J. (1990) Mol. Cell. Biol. 10, 38183823. Zilberstein, A,, Ruggieri, R., Korn, J. H. & Revel, M. (1986) EMBO J . 5,2529 - 2537. zyxwvutsrqpo zyxw zy