The carbohydrate moieties of suppressor IgG-binding factor released by murine T cells

Molecular

Immunology,

Vol. 24, No. 10, pp. 1061-1068,

0161-5890/87$3.00+ 0.00 0 1987Pergamon JournalsLtd

1987

Printed in Great Britain

THE CARBOHYDRATE MOIETIES OF SUPPRESSOR IgG-BINDING FACTOR RELEASED BY MURINE T CELLS ULRICH BLANK,* MARC DA~~RON,ANNIE GALINHA, WOLF HERMAN FRIDMAN and CATHERINE SAUTES? INSERM

U.255, Institut Curie,

26, rue d’ulm,

75005 Paris,

France

(First received 11 February 1987; accepted in revised form 14 April 1987) Abstract-The

carbohydrate moieties of murine IgG-binding factor (IgG-BF) were studied using lectins sequences such as Concanavalin A (Con A), Lens culinaris agglutinin (LcA), and wheat germ agglutinin (WGA), and lectins binding O-glycosylated sequences such as peanut agglutinin (PNA) and Helix pomatia Agglutinin (HpA). Sources of IgG-BF were: (1) supernatants from T,D,, a T binding

N-glycosylated

cell hybridoma constitutively producing IgG-BF, and (2) factor purified by affinity chromatography on rabbit IgG-Sepharose, using T2D4 supernatants or supernatants of alloantigen-activated T cells (ATC)

as starting material. The presence of IgG-BF was assessed by its ability to inhibit secondary anti-sheep red blood cell (SRBC) IgG antibody responses in uirro and to inhibit rosette formation between Fey receptor (Fey R)-positive spleen cells and erythrocytes sensitized with rabbit anti-Forssman IgG antibodies. Fractionation on immobilized lectins showed that IgG-BF: (1) is completely adsorbed by WGA and PNA and partially by Con A, LcA and HpA, and (2) can be eluted from the five different lectins using the competitor sugars. When produced in the presence of tunicamycin, an inhibitor of Nglycosylation, IgG-BF still binds to HpA which has affinity for O-glycosylated results indicate that IgG-BF is a glycoprotein with N- and O-glycosylated

INTRODUCTION

Immunoglobulin-binding factors (IBF) comprise a family of lymphokines which bind to the Fc portion of immunoglobulins (Ig) and play an important role in the regulation of the production of Ig isotypes (Fridman et al., 1986). IBF specific for IgG (IgG-BF) (Fridman and Golstein, 1974; Le Thi Bich-Thuy et al., 1980), for IgE (IgE-BF) (Yodoi’ et al., 1980; Hirashimia et al., 1980) and IgA (IgA-BF) (Yodoi’ et al., 1983; Kiyono et al., 1985), regulate the production of IgG, IgE and IgA respectively. Although only suppressor IgG-BF have been described (Gisler and Fridman, 1976), IgE-BF and IgA-BF may exert enhancing or suppressive properties on the in vitro antibody responses (Yodoi’ et al., 1980; Hirashima et al., 1980; Yodoi et al., 1983; Kiyono et al., 1985). A T cell hybridoma (T, D4) which releases constitutively IgG-BF upon incubation in serum-free medium, was established by the fusion of alloantigen-activated T cells (ATC) and T lymphoma cells (Niauport-Sautis et al., 1979). The production of IgG-BF can also be *Supported by a fellowship from the Institut Scientifique Roussel. TAuthor to whom correspondence should be addressed. Abbreviations: IBF, immunogloblulin-binding factor; FcR, receptors for the Fc portion of immunoglobulins; ATC, alloantigen-activated T cells; pl, isoelectric point; Con A, Concanavalin A; I..cA, Lens culinaris agglutinin; PNA, peanut agglutinin; WGA, wheat germ agglutinin; HpA, Helix pomafia agglutinin; SRBC, sheep red blood cells; aMM, a-methyl-D-mannoside; GlcNAc, N-acetylD-glucosamine; GalNAc, N-acetyl-D-galactosamine; D-Gal, D-gahCtOSe; HAT, hypoxanthin-aminopterinthymidine; PMSF, phenylmethylsulfonyl fluoride.

carbohydrate carbohydrate

chains. These moieties.

induced either in vitro by incubating T,D, cells with IgG or in vivo by growing T, D4 cells in ascites (L&y et al., 1983; Blank et al., 1986~). In vitro treatment of T,D,, cells with IgGl, IgG2 or IgA, results in the production of subclass-specific IgGl-BF, IgG2-BF (Liiwy et al., 1983), or IgA-BF, respectively (Yodoi’ et al., 1983). In vivo induced factor is also present in supernatants of alloantigen-activated T cells (ATC) (Fridman and Goldstein, 1974). Since IgG-BF is synthesized by ATC which express Fey R, and since its productioin is concomitant with a decrease in FcyR expression (NCauport-Saut&s et al., 1986), it has been postulated that IgG-BF may derive from membrane Fey R. The reactivity of IgGBF with the monoclonal anti-FcyR antibody 2.462 (Da&on et al., 1986), and of IgE-BF with a monoclonal anti-FceR antibody (Huff et al., 1984), supports this hypothesis. Membrane FcR of lymphoid cells are surface glycoproteins of M,47-70,000 with a high degree of glycosylation (NCauport-Saut&s et al., 1986). After treatment with neuraminidase (Anderson, 1982) or synthesis in the presence of tunicamycin (Hempstead et al., 1981), FcR have a more uniform mobility in 2-D gels. They can be labeled externally on galactose residues (Mellman and Unkeless, 1980) or internally with radioactive sugars (Green et al., 1985). Four N-linked oligosaccharide chains are present in the 37,000 mol. wt polypeptide of the mouse macrophage FcyR (Green et al., 1985). Previous studies have shown that the M, of IgG-BF is around 35,00&40,000 (Blank et al., 1986~). IgGBF is heterogeneous in charge, the pl of biologically active components being respectively 5.1,6.5, 7.7 and

1061

Uratcn

1062

BLANK

8.4 (Blank et al., 1986~). This wide charge heterogeneity is related to polypeptide differences rather than to glycosylation differences (Blank el al., 198&z,b). Similarly to FcR, IBF are glycoproteins (Neauport-Sautes et al., 1986) and bind to various lectins (Yodoi’ et al., 1982; Sandor et al., 1986; Noro et al., 1986). Interestingly, the glycosylation of IgEBF seems to be directly involved in the biological activities of the factors. IgE-suppressive factor and IgE-potentiating factor differ only in their carbohydrate moieties. Suppressive factor is only Oglycosylated and binds to peanut agglutinin (PNA) but not to Concanavalin A (Con A) and Lens culinaris agglutinin (LcA), whereas potentiating factor is 0- and N-glycosylated and hence binds to the three latter lectins (YodoI et al., 1982). Neuraminidase treatment of potentiating factor results in a loss of its biological activity and the addition of tunicamycin, which prevents N-glycosylation during the production of potentiating factor, leads to the secretion of IgE-BF with suppressive properties (Y’odoi’ et al., 1982). The mechanism(s) by which this differential glycosylation of a unique polypeptide chain occurs have been approached by Ishizaka and colleagues with a T cell hybridoma. It has been shown that suppressive and potentiating IgE-BF can be produced by a single cell which, under the influence of glycosylation enhancing factor (GEF) or glycosylation inhibitory factor (GIF), synthesizes potentiator or suppressor factors, respectively (Ishizaka, 1985). In view of these results, one wonders whether the same correlation between the nature and extent of glycosylation of IgE-BF and their biological effects also holds for other members of the IBF family. In fact, T,D,-derived suppressive IgA-BF is Nglycosylated (Noro et al., 1986). We therefore undertook studies to investigate the glycosylation pattern of IgG-BF produced by T,D, cells. The binding of IgG-BF to five different insolubilized lectins was examined, three leetins, Con A, LcA and wheat germ agglutinin (WGA), having affinity for ~-glycosylated oligosaccharide chains, and two other lectins, peanut agglutinin (PNA), and Helix pomatia agglutinin (HpA), having affinity for 0-glycosylated chains. Crude cell-free supernatants of T,D, hybridoma cells and IgG-BF purified by affinity chromatography on rabbit IgG were used for these studies. The presence of IgG-BF was assessed by its ability to suppress secondary in vitro anti-sheep red blood cells (SRBC) antibody responses and to prevent the binding of IgG to FcyR in a rosette inhibition assay. The results show that murine IgG-BF is highly glycosylated and that, although suppressive, it bears 0- and Nglycosylated carbohydrate chains. MATERIALS

AND

METHODS

Cells IgG-BF were obtained from the T cell hybridoma

e2 d.

T,D, (Neauport-Sautes et al., 1979) and its derived subclone C,TZDd, and from alloreactive ATC (Fridman and Goldstein, 1974). The Fey R-positive T cell line T,D, is a fusion product between BlO.BR anti-BALB/c ATC (H-2k) and the HAT-sensitive FcR-negative thymoma BW-5147 (H-2k). It is maintained continuously in RPM1 1640 medium containing 100 U/ml penicillin-streptomycin, 2 mM L-glutamine and 10% fetal calf serum (FCS) (Gilbco, Paris, France) in a COB-humidified atmosphere at 37°C. ATC are spelenocytes prepared from lethally irradiated BALB/c mice which have been injected 5 days earlier with C3H or C57BL/6 thymocytes (lo8 cells per mouse). Sources of IgC-BF IgG-BF-containing cell-free supernatants were prepared by incubating T2D4 cells or ATC at a final concn of 2 x lo6 cells/ml for 3 h in BSS at 37°C in a CO,-humidified atmosphere. The protease inhibitors PMSF (lo-‘M) and aprotinin (0.1 U/ml) (Sigma, St-Louis, MO) were added to cell-free supernatants. In some experiments, T2D4 cells were cultured in the presence of 2 pg/ml tunicamycin (Sigma) for 24 hr. Affinity-purified IgG-BF were prepared from either T,D, or ATC supernatant cont. 4 times on Amicon PM10 membrane filters (Amicon, Danvers, MA). One vol of purified rabbit IgG coupled at IO mg/mI to Sepharose 4B (Pha~acia, Uppsala, Sweden) was incubated at 4°C for 3 hr or overnight with 2 vols of cont. supernatant, dialysed against 0.02 h4 sodium phosphate buffer, pH 7.0. Beads were washed in the same buffer until the effluent had an O.D.,,, < 0.02. IgG-BF-containing material was eluted with 0.2 M glycine-HCI buffer, pH 2.8, neutralized immediately with 1 M I&H PO, and cont. 10 times as compared to the starting supernatant in Minicon B 15 (Amicon, Danvers, MA). The material was then tested for IgG-BF activity. Sinding of IgG-BF to d@erent lectins LcA, HpA, WGA and PNA lectins, coupled to Agarose beads, were purchased from IBF (Paris, France); Con A lectin coupled to Sepharose was purchased from Sigma. Column procedures were used to investigate the lectin-binding properites of IgG-BF present in crude supernatants whereas batch procedures were used for a~nity-p~fied IgG-BF. Prior to each experiment, beads were washed twice alternatively with about 20 bead vols each time with 20mM sodium acetate buffer, pH 4.0, containing 0.5 M NaCl and 20 mM Tris buffer, pH 9.0, containing 0.5 M NaCl. Washing was pursued with 20-50 vols of 20 mM Hepes buffer, pH 7.2, containing 0.15 h4 NaCl, 1 mM CaCl,, 1 mM MnSO,, and 0.1% Triton X-100. The beads were then equilibrated with lo&150 vols of 20mM Hepes buffer, pH 7.2, containing 0.15 M NaCl, 1 mM CaCl,, and 1 mN supernatants, dialysed MnSO, . Four time-cont. against equilibration buffer, were adsorbed to the

IgG-binding factor carbohydrates different lectins for 12-18 hr, using a column/supernatant ratio of l/6. After collecting the effluent, columns were washed with IO&150 vols of equilibration buffer and elution was carried out with 4 vois of 0.3 M competitor sugar dissolved in equilibration buffer. One ml fractions were collected. Specific sugars were aMM (Sigma) for LcA and Con A lectins, D-Gal (Merck, Damstadt, F.R.G.) for PNA lectin, GlcNAc (Sigma) for WGA lectin and GalNAc (Sigma) for HpA. Batch lectin adsorption of purified IgG-BF was carried out by incubating I vol of lectin with 2~01s of IgG-BF from either TrD, or ATC overnight at 4’C. After washes (see above), elution was made twice with I vol of equilibration buffer containing 0.3 M of the specific sugar. Prior to testing, all materials (supernatants, effluents, and eluates) were dialysed against PBS. Inhibition of IgG antibody production The biologic activity of the different lectin fractions was examined in secondary in oirro anti-SRBC responses (Gisler and Ftidman, 1976). Spleen cells (6-8 x 106) from B,D*F, mice (IRSC-CNRS, Villejuif, France) primed i.p. 7-9 days earlier with 2 x IO’ SRBC, were resuspended in 1 ml RPM1 1640 supplemented with 50 U/ml penicillin-steptomycin, 2 mM L-glutamine, 10% heat-inactivated FCS and I% heat-inactivated horse serum in Falcon plastic tubes (Falcon, Oxnard, CA) under a 5% CO2 humidified atmosphere at 37°C. SRBC (5 x 106), together with material to be tested, were added at day 0 of the cultures. Five days later, IgG plaque-forming cells were detected using an hemolytic plaque assay in liquid medium. Guinea pig serum, absorbed on SRBC, was used as a source of complement. IgG plaques were developed with a rabbit anti-mouse IgG antiserum (Litton, Bionetics, Kensington, MD), at l/l50 final dilution. The percent of suppression was calculated by using as 100% control the numbers of indirect PFC of cultures to which similar vols of PBS wcrc added. More than 30% inhibition of the response was considered as significant. This was verified by the Student r-test. Rosette inhibition assay Material fractionated on lectins was also tested for its ability to inhibit rosette formation by BALB/c splenocytes and SRBC sensitized with rabbit IgG anti-Forsmann antibodies. Prior to the test, the optimal coating conditions for rosette inhibition were determined by using serial dilutions of antibodies. A dilution that gave submaximal rosette formation was chosen. IgG-coated SRBC (1% final concn) were incubated overnight at 4°C on a rotator with the different fractions (l/5 final dilution) in 0.1% BSAcontaining PBS (50 ~1 final vol). Twenty microliters of the mixture were then added to 20~1 of splenocytes resuspended at 5 x 106/m1 in 10% FCScontaining BSS. After contrifugation at 800 rpm (420g) for 2 min at 4”C, pellets were incubated on ice

1063

for 45 min. The cells were gently resuspended and examined for rosette formation. Rosette inhibition was calculated by using, as a 100% control, the proportion of rosettes formed by the same cells in the presence of indicator cells incubated with 0.1% BSA-PBS alone. Twenty per cent inhibition of rosettes was considered as significant.

RESULTS

IgG-BF were fractionated on insolubilized lectins. The distribution of IgG-BF activity was subsequently determined in effluents and eluates. In preliminary experiments, the interference of any residual sugar or of material possibly leaking from the insolubilized lectins and still present after dialysis, was tested. Neither suppressive, nor rosette inhibitory activities were found in effluents or eluates of lectin columns to which buffer alone was applied (data not shown). Presence of‘N-linked carbohydrate moieties in IgG-BF In a first set of experiments, the binding of IgG-BF to WGA which has high affinity for “hybrid type” structures with terminal GluNAc (Cummings and Kornfeld, 1982) was investigated. T, D, serum-free supernatants and IgG-BF purified from these supernatants were passed on insolubilized WGA. Starting materials, effluents and eluatcs recovered with 0.3 M GluNAc were tested for suppressive and IgG-binding activities. As shown in Fig. I, this lectin was able to completely adsorb the biological properties of IgGBF: the effluents were devoid of significant activities whereas eluates significantly suppressed the secondary in oitro IgG anti-SRBC response and significantly inhibited the formation of IgG rosettes. These results suggest that IgG-BF contains N-linked carbohydrate chains. However, WGA also binds certain sequences with terminal neuraminic acid which exist in Oglycosylated molecules (Bhavanandan et al., 1977). The binding of IgG-BF to lectins having affinity for other N-glycosylated sequences was thus investigated next. Con A is specific for biantennary and high mannose type structures (Cummings and Kornfeld, 1982) and LcA is specific for tri- or tetraantennary structures containing internal fucose (Kornfeld et al., 1981). Using either of these lectins, biological activities were recovered in both the 0.3 M rMM eluates and the effluents after fractionation of T,D, supernatants and the IgG-BF on these lectins (Figs 2 and 3). The affinity of IgG-BF, secreted by in cico alloantigen-activated T cells (ATC) was also analysed. Similarly to IgG-BF constitutively secreted by T, D, cells, the suppressive activity of in Git‘oinduced factor was distributed in the effluents and the eluates of immobilized Con A and LcA. In a representative experiment, the Con A effluent and eluate inhibited at l/IO dilution, 30 and 50% of the control response respectively, whereas the starting material inhibited 97% of the indirect plaques at the

ULRICHBLANKet al.

1064

WGA % INHIBITION OF IgG ROSETTE FORMATION

% INHIBITION OF g IgG ;~TIBOC)~ RESP$NSE So

“%%F

I

I

I

I

I

I

I

SN EF EL IgG-BF

I

I

1

I

EF EL

Fig. 1. Fractionation on immobilized WGA of supernatant from T2D4 ceils (SN) and IgG-BF affi~ty-purified from T,D, su~matants. Starting material, effluent (EF), and material eluted with 0.3 M Glu-NAc (EL) were tested for suppression of a secondary in trifro anti-SRBC response (I/IO final dilution) and for inhibition of rosettes formed by murine spleen cells with IgG-sensitized SRBC (l/5 final dilution). In the case of supernatants, four eluate fractions were collected. Values given correspond to the strongest inhibitory fraction obtained as compared to positive controls (7819 + 1708 indirect PFC/106 recovered cells; 46.0% cells forming IgG rosettes).

Con A % INHIBITION OF IgG ANTIBODY RESPONSE

‘%%iTS SN

% INHIBITION OF IgG ROSETTE FORMATION

0

0

r-?++-?

f+-+++

I

I

I

J

EF EL IgG-BF

I

I

I

1

EF EL

Fig. 2. Fractionation on immobilized Con A of supematant from T2D4 cells (SN) and IgG-BF affinity-purified from T2D4 supernatant. Starting material, effluent (EF), and material eluted with 0.3 M aMM (EL), were tested for suppression of a secondary in v&o anti-SRBC response (l/l0 final dilution) and for inhibition of IgG rosette formation (l/S final dilution), In the case of su~matant, four eluate fractions were collected. Values given correspond to the strongest inhibitory fraction obtained as compared to positive controls (2680 * 313 indirect PFC/106 recovered cells; 48.3% cells forming IgG rosettes).

same dilution (data not shown). In the case of LcA, the starting material, effluent and eluate inhibited 42, 82 and 43% of the response respectively (data not shown). These results indicate that IgG-BF is Nglycosylated. In addition, they suggest that the Nlinked oligosaccharide chains are complex and have varible degrees of substitution. Presence of 0-glycosidiraZly chains in IgG-BF

Iinked

carbohydrate

Two lectins were used: HpA and PNA. HpA binds preferentially terminal GalNAc residues, Oglycosidically linked to the protein core (Goldstein

and Hayes, 1978). When TzDs supernatant or affinity-purified IgG-BF preparations were fractionated on HpA-Agarose, suppressive and rosette inhibitory activities were found in the eluates recovered with 0.3 M GalNAc (Fig. 4). Effluents were suppressive but had no significant effect on rosette formation. Since HpA binds not only to GalNAc but also to GluNAC present in ~-glycosidiGaIly linked sugars, although with a much lower affinity (Goldstein and Hayes, 1978), studies using tunicamycin as inhibitor for N-glycosylation were undertaken to determine if the interaction of IgG-BF with HpA occurs only via O-glycos~di~ally linked sugars. T,D,

IgG-binding

factor

LCA

p%E:Ts

% INHIBITION OF o lgG i44TIBOD$ RESP$NSE *g I

SN

I

I

I

1

I 1

I

EF EL IgG-BF

I

I

EF EL

Fig. 3. Fractionation on immobilized LcA of T,D, supernatant (SN) and IgG-BF, affinity purified from T,D, supernatant. Starting material, effluent (EF) and 0.3 M aMM eluates (EL) were tested for suppression of a secondary in vitro anti-SRBC response (l/IO final dilution). In the case of supernatants, four eluate fractions were collected. Values given correspond to the strongest inhibitory fraction obtained as compared to positive controls (4658 k 1205 indirect PFC/106 recovered cells).

cells were cultured for 24 hr in the presence of 2 pg/ml tunicamycin. Supernatants prepared from these cells were still biologically active and inhibited 66% of secondary in vitro anti-SRBC responses. Since inhibition of N-glycosylation by tunicamycin may not be complete, supernatants were first fractionated on WGA-Agarose in order to deplete completely any residual factor with N-glycosidally linked chains. By contrast with the results obtained above, the effluent of the WGA column exhibited significant suppressive activity on IgG antibody responses indicating that tunicamycin treatment was effective (Fig. 5). This material was further fractionated on HpA. The eluate was suppressive showing that IgG-BF

carbohydrates

1065

deprived of N-glycosylated chains still binds to HpA. Therefore these data suggest that: (1) IgG-BF contains 0-glycosylated carbohydrate chains and (2) N-glycosylated chains are not necessary for the biological activities of the factor. The binding of the IgG-BF to PNA, which binds preferentially the sequence GalaGalNAc in Oglycosidically linked carbohydrate chains (Nicholson and Irimura, 1984), was also assessed. When T,D, supernatant and affinity-purified IgG-BF were fractionated on immobilized PNA, both suppressive and rosette inhibitory activities were adsorbed. As shown in Fig. 6, the activities were recovered in the 0.3 M D-Gal eluates in both cases. Similar results were obtained when IgG-BF from ATC was analysed (data not shown). Therefore these results confirm that IgG-BF contain carbohydrate chains linked Oglycosidically to the serine and/or threonine residues of the polypeptide chains. DISCUSSION

The present study shows that murine IgG-BF released by hybridoma T cells and from activated T cells are glycosylated molecules binding to five different lectins. Binding to WGA and PNA was complete. On LcA, Con A and HpA, biological activities were detectable in the effluents. Immunosuppressive and rosette inhibitory activities could be recovered in the eluates of all five lectins using the appropriate competitor sugars. Several conclusions can be drawn from these data. The binding of IgG-BF to these five lectins indicates that IgG-BF is N- and 0 -glycosylated. The N-linked carbohydrate chains seem complex like those of most glycoproteins (Cummings and Kornfeld, 1982; Neel et al., 1985). The adsorption to WGA suggests that IgG-BF probably bear N-glycosylated “hybrid type”

HPA %

PRODUCTS

INHIBITION OF IgG ANTIBODY RESPONSE

ADDED II

% INHIBITION OF IgG ROSElTE FORMATION II

SN EF EL IgG-BF EF EL

Fig. 4. Fractionation on immobilized HpA of supernatant from T,D, cells (SN) cont. 4 times and IgG-BF alfinity-purified from these supernatants. Starting material, effluent (EF) and material eluted with 0.3 M GalNAc (EL) were tested for suppression of a secondary in vitro anti-SRBC response (l/l0 final dilution) and for inhibition of IgG rosette formation (l/S final dilution). In the case of supernatant, four eluate fractions were collected. Values given correspond to the strongest inhibitory fraction obtained as compared to positive controls (4658 f 1205 indirect PFC/106 recovered cells; 44.0% cells forming IgG rosettes).

ULRICH BLANK et al.

1066

HpA (Tunicamycin) % INHIBITION OF IgG ANTIBODY RESPONSE

pR%%Ts I

Start. mat.

I

EF EL

Fig. 5. Fractionation on immobilized HpA of IgG-BF produced in the presence of tunicamycin. Supernatant from T2D4 cells cultured with tunicamycin was first applied on WGA columns in order to remove any remaining Nglycosylated molecules. The effluent from such a column was then applied on HpA columns. This material (Start. mat.), 0.3 M GalNAc eluate (EL) and eflluent from HpA (EF) were tested for suppressive activity on secondary in oirro IgG-anti-SRBC responses (I/l0 final dilution). Values given correspond to percent inhibition of IgG-anti-SRBC response as compared to positive controls (12051 f 229 indirect PFC/106 recovered cells). Supernatants of T,D, cells cultured in the presence of tunicamycin inhibited 66% of the response.

carbohydrate chains containing terminal GluNAc linked to the mannose residue (Cummings and Kornfeld, 1982). The partial adsorption to Con A indicates that some oligosaccharide chains have biantennary structures of that IgG-BF preparations contain metabolic intermediates with polymannosyl structures formed during the process of glycosylation (Hubbard and Ivatt, 1981). Finally, binding of IgGBF to LcA indicates that some biologically active molecules contain tri- or tetraantennary structures with internal fucose residues (Kornfeld et al., 1981). O-linked carbohydrate chains are also present in IgG-BF, since IgG-BF binds to HpA, even when it is deprived of N-linked sequences, and binds also to

PNA which has high affinity of 0-glycosydically linked galactose (Nicholson and Irimura, 1984). Although no quantitative conclusions can be drawn from the biological assays used herein, the existence of significant suppressive and rosette inhibitory activities in the effluents of Con A, LcA and HpA, suggests that glycosylation of IgG-BF may be heterogeneous. In the case of Con A and LcA it probably indicates variable degrees of substitution of the N-glycosydically linked chains, as already observed for human MHC class II molecules (Neel et al., 1985). Repartition of the suppressive activity in the effluents and the eluates of HpA columns reflects that IgG-BF components may not all be Oglycosylated. Interestingly the effuents of HpA columns exerted no effect on rosette formation. They might contain the most basic component (pZ8.4) of IgG-BF which is suppressive but does not inhibit rosette formation (Blank et al., 1986~). If some of the IgG-BF molecules are not 0-glycosylated, the complete binding of the factor to PNA seems surprising. This lectin could bind N-linked chains via galactose, although with a lower affinity than to the sequence Gal!GalNAc present in O-linked chains (Goldstein and Hayes, 1978). Since IgG-BF preparations contain four biological active components differing in their apparent pZ, the present data raise the question whether they all are N- and 0-glycosylated. In fact it seems to be the case since all the suppressive and IgG-binding activities were adsorbed onto WGA and PNA. On the other hand, analysis by NEHGE of eluates of PNA columns revealed the same charge heterogeneity as the starting material (pZ5.1, 6.2, 7.6 and 8.4). The present findings indicate that suppressive IgG-BF is not only 0-glycosylated but also Nglycosylated. 0-glycosylation of suppressive human IgG-BF (Sandor et al., 1986) and of suppressive and

PNA % INHIBITION OF IgG ROSETIE FORMATION

% INHIBITION OF IgG ANTIBODY RESPONSE

p%!!Fs

O-+-+-9+ SN

I

I

I

I

EF EL IgG-BF

I

1

I

EF EL

Fig. 6. Fractionation on immobilized PNA of supematant from T,D, cells (SN) and IgG-BF, affinity-purified from these supematants. Starting material, effluent (EF), and material eluted with 0.3 M D-Gal (EL) were tested for suppression of a secondary in vitro anti-SRBC response (l/IO final dilution) and for inhibition of IgG rosette formation (l/5 final dilution). In the case of supernatants, four eluate fractions were collected. Values given correspond to the strongest inhibitory fraction obtained, as compared to positive controls (7819 k 1708 indirect PFC/106 recovered cells; 46.0% cells forming IgG rosettes).

IgG-binding factor carbohydrates

IgE-BF (Yodoi’ et al., 1982), was also In contrast to IgE-BF where Nglycosylation is characteristic of potentiating activity, N-linked carbohydrates can clearly be detected in suppressive IgG-BF. Similar findings have been obtained with suppressive IgA-BF secreted by murine T cells or by T,D, cells upon induction with IgA (Noro et al., 1986) and with the IgE suppressive factor of Suemura et al. (1981). The reasons for these differences in glycosylation among the various IBF are still unclear. IgG-BF and IgE-BF used for these studies come from two distinct subsets of T cells, Lyt2+ (Ntauport-Sautes et al., 1979) and Lytl + (Ishizaka, 1985) respectively, which may have different glycosylation processes. Another possibility is that the different IBF correspond to totally unrelated proteins. In fact, the recent cloning of IgE-BF has led to the isolation of cDNA having homology with retroviral sequences (Moore et al., 1986). By contrast, cDNA encoding Fey R, has no homology with the latter sequences (Hibbs et al., 1986) but encodes a protein with four putative N-glycosylation sites (Green et al., 1985). The present work also extends and confirms previous observations concerning the role of carbohydrates in the biological effects of IgG-BF. First, we showed that IgG-BF deprived on N-glycosylated chains by synthesis in the presence of tunicamycin is biologically active. Second, previous results have indicated that unglycosylated IgG-BF synthesized by in vitro translation of T,D, mRNA in a rabbit reticulocyte lysate system inhibits strongly secondary in vitro anti-SRBC responses and has affinity for IgG (Blank et al., 1986a, bj. Finally, neuraminidase treatment of IgG-BF significantly alters neither immunosuppressive properties nor binding to IgG (U.B., unpublished data). It seems, therefore, that in contrast to IgE-BF, carbohydrates do not play a significant role in the biological activity of IgG-BF. These differences may reflect that IgE-BF and IgGBF interact with different target molecules on B cells, most probably membrane IgE (Uede et al., 1984) and IgG (Fridman et al., 1986), respectively. Thus, from presently available data, it appears that the relationship between the glycosylation pattern and the biological activity described for IgE-BF cannot be extended to every other member of the IBF family. potentiating observed.

Acknowledgements-We gratefully acknowledge Drs D. Neel and Y. Goussault (INSERM U.180, Paris, France) for their helpful information, the members of U.255 for stimulating discussions and C. Amsellem for typing this manuscript. This work was supported by grants from the Institut National de la Sante et-de la Recherche Medicale (INSERM) and the Centre National de la Recherche Scientifique (CNRS grant No. 960138). REFERENCES

Anderson C. L. (1982) Isolation of the receptor for IgG from a human monocyte cell line (U937) and from human peripheral blood monocytes. J. exp. Med. 156,17941806.

1067

Bhavanandan V. P., Umemoto J., Banks J. R. and Davidson E. A. (1977) Isolation and partial characterization of sialoglycopeptides produced by a murine melanoma. Biochemistry 16, 442-37. Blank U., Fridman W. H., Da&on M., Galinha A., Moncuit J. and Ntauport-Sautes C. (1986~) Size and charge heterogeneity of murine IgG-binding factors (IgG-BF). J. Immun. 136, 2975-2982. Blank U., Fridman W. H., Da&on M., Galinha A., Moncuit J. and Neauport-Sautes C. (19866) Charge heterogeneity of murine IgG-BF. In Protides of the Biological Fluids (Edited by Peters H.), 34th Coll., pp. 545-548. Pergamon Press, Oxford. Cummings R. D. and Komfeld S. (1982) Fractionation of asparagin linked oligosaccharides by serial lectin-agarose affinity chromatography. J. biol. Chem. 257, 11235-l 1240. Da&on M., Neauport-Sautes C., Blank U. and Fridman W. H. (1986) 2.462, a monoclonal antibody to macrophage Fc y receptors, reacts with murine T cell Fey receptors and IgG-binding factors. Eur. J. Immun. 16, 1545-1550. Fridman W. H. and Golstein P. (1974) Immunoglobulinbinding factor present on and produced by thymusprocessed lymphocytes (T cells). Cell. Immun. 11, 442-455. Fridman W. H., Teillaud J. L., Amigorena S., Da&on M., Blank U. and Nbauport-Sautes C. (1986) The isotypic circuit: immunoglobulins, Fc receptors and immunoglobulin-binding factors. Znt. Rev. Zmmun. 2. 19-38. Giiler R. and Fridman W. H. (1976) Inhibition of the in vitro 19s’ and 7S antibody response by immunoglobulinbinding factor (IBF) from alloantigen-activated T cells. Cell. Immun. 23, 99-107. Goldstein I. J. and Hayes C. E. (1978) The lectins: carbohydrate-binding proteins of plant and animals. Adu. Carbohydr. Chem. Biochem. 35, 127-340. Green S. A., Plutner H. and Mellman J. (1985) Biosynthesis and intracellular transport of the mouse macrophage Fc receptor. J. biol. Chem. 260, 9867-9874. Hempstead B. L., Parker C. W. and Kulczycki A., Jr (1981) The cell surface receptor for immunoglobulin E. Effect of tunicamycin on molecular properties of receptor from rat basophilic leukemia cells. J. biol. Chem. 256, - 10717-10723. Hibbs M. L., Walker J. D., Kirszbaum L., Pietersz G. A., Decon N. J., Chambers G. W., McKenzie J. F. C. and Hogarth P. M. (1986) The murine Fc receptor for immunoglobulin: purification, partial amino acid sequence, and isolation of cDNA clones. Proc. natn. Acad. Sci. U.S.A. 83, 698&6984. Hirashima M., Yodoi’ J. and Ishizaka K. (1980) Regulatory role of IgE-binding factors from rat T lymphocytes III IgE-specific suppressive factor with IgE-binding activity. J. Immun. 125, 1442-1448. Hubbard S. C. and Ivatt R. J. (1981) Synthesis and processing of asparagine linked oligosaccharides. A. Rev. Biochem. 50, 555-583. Huff T. F., Yodoi’ J., Uede T. and Ishizaka K. (1984) Presence of an antigenic determinant common to rat IgE-potentiating factor, IgE-suppressor factor, and Fct receptors on T and B lymphocytes. J. Immun. 132, 406412. Ishizaka K. (1985) Presidential address at the American Association of Immunologists: twenty years with IgE: from the identification of IgE to regulatory factors for the IgE response J. Zmmun. 135, i-x. Kiyono H., Mosteller-Barnum L. M., Pitts A. M., Williamson S. I., Michalek S. M. and McGhee J. R. (1985) Isotype-specific immunoregulation. IgA-binding factors produced by Fc (Yreceptor positive T cell hybridomas regulate IgA responses. J. exp. Med. 161, 731-747. Komfeld K., Reitman M. L. and Kornfeld R. (1981) The

1068

ULRXH BL.ANK et a/.

Nicholson G. L. and Irimura T. (1984) Estimating lycocarbohydrate-binding specificity of pea and lentil lectinsproteins carbohydrate chain structures by lectin reac: fucose is an important determinant. J. biol. Chem. 256, iivities in polya&ylamide gels. Biol. Cell. 51, 157-164. 6633-6640. Noro N.. Adachi M.. Yasuda K., Masuda T. and Yodoi’ J. Le Thi Bich-Thuy, Samarut C., Brochier J., Fridman W. H. (1986) Murine IgA binding factors (IgA-BF) suppressing and Revillard J. P. (1980) Suppression of mitogenIgA production: characterization and target specificity of induced peripheral B cell differentiation by soluble Fey receptors released from lymphocytes. Eur. J. Immun. 10, IgA-BF. J. Immun. 136, 291&2916. 894-898. Sandor M., Erdei A., Blank U., Nkauport-Sautis C., Lijwy I., Brezin C., Ntauport-Sautes C., Thtze J. and Fridman W. H. and Gergely J. (1986) The role of Fridman W. H. (1983) Isotype-specific regulation of carbohydrates in the IgG-BF and suppressive activities of antibody production: T cell hybrids can be selectively shed human Fc receptors. Annls Inst. Pasteurllmmun. induced to produce IgGl and IgG2 subclass-specific 137D, 77-89. suppressive immunoglobulin-binding factors. Proc. natn. Seumura M., Shiho O., Deguchi H., Yamamura Y., Acad. Sci. U.S.A. 80, 2323-2327. Biittcher I. and Kishimoto T. (1981) Characterization and Mellman 1. S. and Unkeless J. C. (1980) Purification of a isolation of IgE class-specific suppr&sor factor (IgE-TsF). functional mouse Fc receptor through the use of a I. The oresence of the bindine sites for IpE and of the H-2 monoclonal antibody. J. e$. Med. 152, 1048-1069. gene products in IgE-TsF. i Immun. 127, 465471. Moore K.. Jardieu P.. Mietz J. M.. Troustine M. L.. Kuff Uede T., Huff T. F. and Ishizaka K. (1984) Suppression of E. L., ishizaka K. and Martens C. L. (1986) dodent IgE synthesis in mouse plasma cells and B cells by rat IgE-binding factor genes are members of an endogenIgE-suppressive factor. J. Immun. 133, 803-808. ous retrovirus-like gene family. J. Immun. 136, Yodoi. J.: kdachi M., Teshigawara K., Miyama-Inaba M., 4283-4290.24. Masuda T. and Fridman W. H. (1983) T cell hvbridomas coexpressing Fc receptors (FcR) ‘for different iiotypes. II. N&auport-Sautis C., Da&on M., Teillaud J. L., Blank U. IgA-induced formation of suppressive IgA-binding facand Fridman W. H. (1986) The occurrence, structural and tor(s) by a murine T hybridoma bearing FcyR and FcclR. functional properties of immunoglobulin Fc receptors on murine neoplastic cells. Int. Rev. Immun. 1, 237-271. J. Immun. 131, 303-310. Niauport-Sa&s C., Rabourdin-Combe C. and Fridman Yodoi’ J., Hirashima M. and Ishizaka K. (1980) Regulatory W. H. (1979) T-cell hybrids bear FCY receDtors and secrete role of IgE-binding factors from rat T lymphocytes. suppressor immunogiobulin-binding factor. Nature 277, II. Glycoprotein nature and source of IgE potentiating 656658. factor: J.-Immun. 125, 143&1441. _ Neel D., Giner M., Merlu B., Turmel P., Goussault Y. and Yodoi’ J.. Hirashima M. and Ishizaka K. (1982) Renulatorv Charron D. J. (1985) Analysis of HLA-DR antigen role 0; IgE-binding factors from rat T‘lymphoc$es. Vi. glycosylation by serial lectin affinity chromatography. The carbohydrate moieties in IgE-potentiating factors Molec. Immun. 22, 1053-1059. and IgE-suppressive factors. J. Immun. 128, 289-295.