ete atinrtnkeidne Cid has on ER afte PA wed me Concanavalin A Reaction with Human Normal Immunoglobulin G ready described, except that 10, 30, or 60 mg of concanavalin A were added and Myeloma Immunoglobulin G per milliliter of each serum. Abstract. Concanavalin A precipitated less than 5 percent of immunoglobulin G from human serum. It reacted with all of 42 myeloma serums of the immunoglobulin G type tested, but no more than approximately 50 percent of the total myeloma protein was ever precipitated. The fact that not all of the protein was precipitated and that the amounts precipitated varied from serum to serum may be interpreted as demonstrating heterogeneity of the carbohydrate in these myeloma proteins. Other glycoproteins precipitated by concanavalin A were iden- fractions were analyzed electrophoret- tified, and subsequently separated from concanavalin A by chromatography. Concanavalin A, the globulin from jack-bean meal, reacts with serum gly- coproteins (7, 2}, presumably because of its specificity for their nonreducing mannosyl- and N-acetylglucosaminy] end groups (3). Harris and Robson demonstrated that crude jack-bean meal extracts precipitated oa,-macroglobulin, yi-macroglobulin (IgM), @-lipoprotein, and ceruloplasmin; as judged by elec- trophoresis haptoglobins were also precipitated (2). Since no reaction with 7S y-globulin (IgG) was noted, I attempted to use concanavalin A to sep- arate IgM from IgG. I now present data demonstrating that, depending on the serum, variable amounts of IgG from normal and myeloma serums precipitated with concanavalin A. Concanavalin A was prepared by a modification of Agrawal and Goldstein’s procedure (4). The crude saline extract of commercial, defatted jack- bean meal wasclarified by centrifugation at 25,000g (0°C) for 1 hour. After removal of the lipid layer, the supernatant was applied directly to a Sephadex G-75 column. Subsequent procedures were as described (4). The purified product was dialyzed against 0.14 phosphate buffer, pH 6.2, and kept in small vials at —70°C to prevent the canavalin A with glycoproteins. Components in the dissociated precipitate, confirmed by immunodiffusion tests with specific antiserums or by specific biological activities, were IgM, IgA, aj1- antitrypsin, qo-macroglobulin, transferrin, 8.-lipoprotein, f2-glycoprotein, haptoglobin, ceruloplasmin, group-specific components, the serum inhibitor of C’l esterase (5) and all the components of complement. Although, by immunoelectrophoresis, no IgG was observed in the dissociated precipitate, both immunodiffusion and Gm typing (6) showed that a small serum, the quantity of IgG precipitated by concanavalin A varied from approximately 50 percent to mere traces of the total present. Immunoelectro- phoretic patterns of the most reactive, an IgG myeloma serum of heavy-chain subtype y3, are shown in Fig. 1. Our limited data do not yet permit correla- tion of heavy-chain subtype with reactivity towards concanavalin A. In all reactive serums, the dissociated myeloma protein showed the same elec- trophoretic mobility as the myeloma protein of the original serum. Increasing the ratio of concanavalin to serum above 30 mg/ml did notsignifi- cantly increase the amount of myeloma protein precipitated. If one accepts the theory that monoclonal myeloma pro- tein in an individual serum is homo- geneous with respect to amino acid sequence, then all molecules of mye- wliin carbohydrate (7) or (ii) differences in availability of IgG carbohydrate for reaction with concanavalian A as a result of conformational differences in IgG molecules with different amino acid sequences. Since homogeneous IgG of both reactive and nonreactive types would facilitate testing of these hypo- theses, 42 IgG myeloma serums were examined for reactivity with concanavalin A (8). Precipitation, washing, and dissociation were performedas al- conformational possibilities in equilibrium. Carbohydrate chains attached that the carbohydrate chains are identi- cal and are attached at the same position in the peptide chain. The partial precipitation of IgG myeloma proteins by concanavalin A, demonstrates existence of some type of heterogeneity of each IgG myeloma carbohydrate. The data further suggest that a given se- quence of amino acids either does not uniquely determine the structure, or does not uniquely determine the location, of the carbohydrate attached to the peptide chain. These findings ex- tend, by entirely different techniques, those reported of carbohydrate heterogeneity in an IgA myeloma globulin (7). , Despite the observation that small amounts of IgG are precipitated by con- canavalin A, the preparation of IgM is cubation at 4°C prior to centrifugation gave only slightly increased yields of glycoproteins. The washed concanaval- 8 DECEMBER 1967 ically. Depending on the IgG myeloma to these molecules should show similar reactivity with concanavalin A provided These data may beattributed to (i) the known heterogeneity of immunoglob- fuged at 4°C and then washed three times with cold saline. Overnight in- coside to prevent recombination of con- nonprecipitable loma protein have the same range of 7 to 8 mg of concanavalin A were added to 1 ml of serum, and the mixture was incubated at 37°C for 1 hour. The flocculent precipitate was centri- terial were performed in agarose gels which contained 0.25M methyl e-D-glu- and amount (<5 percent) of IgG had precipitated with the concanavalin A. precipitation noted on storage at 4°C. To precipitate serum glycoproteins, in-glycoprotein complexes were dissociated at 37°C with 1 ml of saline containing 0.5M methyl! a-D-glucoside. Immunoelectrophoretic and immunodiffusion tests on the dissociated ma- Precipitable Fig. 1. Immunoelectrophoresis of precipitable (top) and nonprecipitable (bottom) fractions obtained from reaction of concanavalin A with serum of a patient with IgG myeloma of heavy-chain subtype ys. Both fractions are diluted 1 to 3 with respect to original serum. The antiserum is horse antiserum to human IgG. greatly facilitated by a preliminary precipitation of the serum with concanavalin A. The chromatographic behavior of concanavalin A makes it readily separable from IgM and from most other serum glycoproteins. When the dissociated mixture of concanavalian A and glycoproteins is applied to diethylaminoethyl Sephadex at ionic strength above 0.02, in tris buffer at pH 8.0. the coacanavalian A is not bound. A com1325