A copper-binding immunoglobulin from a myeloma patient. Purification, identification, and physical characterization.

A copper.protein complex present in the serum of a hypercupremic myeloma patient has been purified to homogeneity using gel filtration, DEAE-cellulose chromatography, and concanavalin A/Sepharose affinity chromatography. Immunoelectrophoresis and hemagglutination inhibition tests showed the copper-bound protein to be an IgG1-type immunoglobulin with lambda light chains. The immunoglobulin is of normal molecular weight (150,000) with normal size light and heavy chains (28,000 and 56,000, respectively). The carbohydrate portion of the molecule appears to be abnormal in that it interacts with concanavalin A, whereas most immunoglobulins of the gammaG-type do not. The copper in the native copper.IgG complex is in an EPR-indeterminable valence state. Copper was efficiently removed from the copper.IgG complex by dialysis against 0.1 M potassium cyanide. The apo-IgG was separated from the copper.cyanide complex by gel filtration. The copper complex was reconstituted by equilibrating the apo-IgG with 7.7 muM cupric ions.

A copper l immunoglobulin complex isolated from a myeloma patient with hypercupremia has been investigated in an attempt to localize and characterize the copper-binding site. Cleavage of the purified IgGl type immunoglobulin with papain yielded peptides which were identified as Fab and F, fragments by immunodiffusion. The fragments were separated by DEAE-cellulose chromatography and the Fab fragment then was purified to homogeneity by chromatography on Sephadex G-100 and protein A-Sepharose.
Copper is bound only to the F,b fragment.
In contrast, the F, fragment, but not the F,b fragment, binds to concanavalin A, suggesting that the carbohydrate moiety on the F, fragment is not involved in the binding of copper. Light and heavy chains were isolated by gel filtration of the products formed upon reaction of the immunoglobulin with mercaptoethanol and iodoacetamide. Acid hydrolysates of light chain, heavy chain, and Fab fragment have been analyzed for amino acid content.
The copper l Fab complex, like the copper*IgG complex, is colorless with no detectable visible absorption peak. Dialysis of the copper 0 Fab against cyanide yields apo-F,b; reconstitution with cupric ions yields a complex with 1 copper atom/F,b fragment. The isolated and reconstituted copper l F,b complexes are EPR-nondetectable before and after treatment with ferricyanide. Comparison of the copper l Fab and apo-F,b in terms of fluorescence spectra, circular dichroism spectra, and immunological properties revealed no structural differences due to copper binding. Reaction with dithiodipyridine demonstrates the presence of one exposed sulfhydryl group in apo-F,b but not in isolated or reconstituted copper l F,h complexes. The dithiodipyridinereacted apo-F,b no longer binds copper, indicating that this sulfhydryl group might be involved in copper binding. A preliminary experiment indicates that the involved sulfhydryl is labeled upon reaction with [14C]iodoacetamide.
In 1976, Lewis et al. (1) described a patient with multiple myeloma and hypercupremia.
Copper balance and radioisotopic studies suggested a normal gastrointestinal and renal threshold for copper but indicated a specific and tight binding between copper and a large molecular weight serum component. This  y-globulin fraction on cellulose acetate electrophoresis, thus distinguishing it from albumin and ceruloplasmin. In a subsequent report (2), we described the isolation of this copper-binding component and identified it as an IgGI' (Gm f) type immunoglobulin with X light chains. The copper-binding immunoglobulin was shown to be of normal molecular weight (150,000) with normal size light and heavy chains. The copper of the immunoglobulin complex exhibited no EPR signal or visible absorbance spectrum. The immunoglobulin was shown to bind 2 atoms of copper which could be removed by dialysis against cyanide. Incubation of the apo-IgG with cupric copper fully reconstituted the copper. immunoglobulin complex.
In the present report, the copper. immunoglobulin complex is further investigated.
We present evidence that 1 atom of EPR-nondetectable copper is located on each Fab fragment and that a thiol group is involved in the binding of copper.  Weber and Osborn (3). One gel was stained with Coomassie brilliant blue to localize protein bands. These protein bands were subsequently excised and radioactivity counted in a liquid scintillation counter.
The second gel was immediately frozen in dry ice and cut into l-mm slices with a Mickle gel slicer. The gel slices were placed in 5 ml of counting fluid (9) gave three peptide peaks. SDS disc gel electrophoresis indicated that the second peak is the heavy chain derivative, the third peak is the light chain derivative, and the first peak is incompletely reacted immunoglobulin.
The third peak was identified as light chain derivative by immunodiffusion against light chain antiserum.

Preparation of Antibodies to the Copper. Fnh Fragment
Male New Zealand white rabbits, weighing 5 to 6 kg, were used for immunization.
A homogeneous copper. Fah solution (approximately 1.0 ml containing 6 mg) was mixed with an equal volume of complete Freund's adjuvant. A paste was prepared by repeatedly passing the mixture between two syringes as described by Chase (11). One milliliter of the paste was injected subcutaneously in four or five spots on the upper back of each of two rabbits.
One week later, a booster of the same concentration of copper .Fnb, this time in incomplete Freund's adjuvant, was administered in an identical manner. Five davs following the booster iniection, approximately 40 ml of blood wis taken fro& the marginal v"ein of an &r of each rabbit and allowed to clot. After centrifugation at 2000 x g for 10 min, the supernatant serum was removed with a pipette.

The antibodies
were partially purified from the serum by several ammonium sulfate precipitation steps. Ten milliliters of saturated ammonium sulfate were added to 20 ml of serum. After 10 min at room temperature, the precipitated immunoglobulins were sedimented by centrifugation at 20,000 x g for 15 min. The supernatant fraction was discarded and the pellet resuspended in 10 ml of water. This process of ammonium sulfate precipitation, centrifugation, and resuspension of the pellet was repeated two additional times. The pellet was then resuspended in water and dialyzed extensively against 0.01 M potassium phosphate buffer, pH 8.0.

AND DISCUSSION
Localization of the Copper on the Fah Fragment-Incubation of the copper. immunoglobulin complex with the proteolytic enzyme papain resulted in cleavage of the molecule into Fah and F, fragments, as well as smaller peptides. DEAEcellulose chromatography effectively separated the hydrolysate into three peptide fractions (Fig. 1). All of the copper was associated with the fraction that was washed through the column with the equilibration buffer without binding (Fraction I). A small amount of peptide with no copper (Fraction II) was eluted with a linear gradient (0 to 0.2 M) of NaCl, and a much larger amount of 280 nm absorbing material (Fraction III) was liberated when the column was stripped with buffer containing 1.0 M NaCl. Electrophoresis on polyacrylamide disc gels showed that the copper. peptide Fraction I (Fig. 2.4) and the non-coppercontaining Fraction II (Fig. 2C) together accounted for all of the peptide bands exhibited by the total hydrolysate ( Fig.  2B). Although each fraction showed multiple bands on disc gels, no cross-contamination between the two fractions could be detected. No peptide bands were detected when Fraction Polyacrylamide disc gel electrophoresis of the papain hydrolysate fractions from the DEAE-cellulose chromatography (see text for details). The gels were formed and run according to the procedures of Brewer and Ashworth (12). Gels were stained with Coomassie brilliant blue. Fraction I is shown in Geld and Fraction II is shown in Gel C. The total hydrolysate is shown in Gel B. Fraction III showed no protein bands when stained with Coomassie brilliant blue (gel not shown). All of the protein bands in the total hydrolysate are found in Gels A or C.
III was electrophoresed and stained in polyacrylamide disc gels, suggesting that this fraction probably consists of very small polypeptides.
Whereas the copper peptide (Fraction I) gave negative precipitin tests with concanavalin A, both of the copperless fractions (II and III) gave positive tests, suggesting that they both contain carbohydrate residues. Sequential chromatography of the copper -fragment on columns of Sephadex G-100 and protein A-Sepharose yielded a fraction that appeared homogeneous.
The purified sample showed a single band upon electrophoresis at pH 8.6 on 7.5% polyacrylamide gel (Fig. 3A). Upon electrophoresis on SDS-polyacrylamide gels in the presence of Z-mercaptoethanol, this same protein fraction exhibited two peptide bands (Fig.  3B), of approximately 28,000 and 33,000 daltons, respectively. Identity of the partially purified fragments produced by papain treatment was accomplished by immunodiffusion against anti-human F.b and anti-human F, antisera. Fig. 4A shows that a precipitin line of complete identity with the copper.immunoglobulin complex ( Wells 1 and 4) was obtained when the copper.peptide (Fraction I, Well 2) was allowed to diffuse against anti-human Fab antiserum (center well). No precipitin band was obtained when the copperless Fractions II ( Well 3) and III ( Well 5) were allowed to diffuse against the same antiserum. Fig. 4B shows that a precipitin line of complete identity with the copper=immunoglobulin complex ( Wells 1 and 4) is observed when Fraction II ( Well 3) was allowed to diffuse against anti-human F, antiserum ( center well). Neither the copper. peptide (Fraction I, Well 2) nor Fraction III (Well 5) formed precipitin bands when allowed to diffuse against this same antiserum.
The results of these immunodiffusion tests suggest that the copper. peptide (Fraction I) is the antibody-combining (Fab) fragment of the copper-binding immunoglobulin.
The copperless Fraction II probably constitutes the F, fragment of the immunoglobulin.
Fraction III probably consists of small peptides derived from the F, fragment of the copper-binding immunoglobulin.
The observation that the F, fragment formed a precipitate with concanavalin A, whereas the copper. Fab complex did not,   suggested that the reactive carbohydrate residues were not located on the Fa,, fragment and were not involved in the binding of copper. Likewise, the affinity of protein A-Sepharose for the copper. immunoglobulin complex but not for the copper. Fab complex indicated that the protein A-reactive portion of the immunoglobulin resided on the F, fragment. These data very strongly suggest that the abnormal copper affinity is due to the unique primary structure of the F,b fragment of the immunoglobulin.
Characterization of the Copper. Fab Complex--The molecular weight of the copper. Fah fragment was estimated from its elution volume from a Sephadex G-100 column. A molecular weight of 50,000 was obtained for this fragment from a standard curve plot of log (molecular weight) versus V,/VO. The copper. Fab complex contained 1.38 atoms of copper/ 50,000 daltons of protein. The copper. Fab exhibited two major peptide bands (Fig. 3B) upon electrophoresis on SDS-polyacrylamide gel in the presence of 2-mercaptoethanol. The molecular weights of these peptides were estimated to be 28,000 and 33,000, which presumably correspond to the presence of light chain and the Fd portion of heavy chain, respectively. The value of 28,000 is the same value obtained for the light chain by electrophoresis of the copper. immunoglobulin in the presence of SDS and mercaptoethanol (2). A minor protein band of approximately 17,000 daltons was seen on overloaded gels. This band was not observed when homogeneous light chain or copper a immunoglobulin were electrophoresed in this system and was presumed to be a degradation product of the Fd fragment of the heavy chain.
The UV and visible absorption spectra of the copper. Fab fragment are shown in Fig. 5A. No absorption peak was detected in the visible region of the spectrum at the 80.8 pM concentration employed. Under these conditions, absorption with a molar extinction of 60 would have been detectable. It should be pointed out that a number of copper complexes have absorptivities less than this in the wavelength range reported here. The UV spectrum showed a single sharp ab- The UV CD spectra of the copper. Fab fragment are shown in Fig. 6A. The far UV spectrum showed a large negative Cotton effect at 218 nm. Negative Cotton effects near this region as well as a positive band at 200 nm have been shown to be indicative of /3 structure (13)(14)(15). In the near UV region, a positive Cotton effect was observed at 285 and 290 nm. Although the effects of other aromatic side chains cannot be fully excluded, bands near this region have been attributed to tryptophan (16, 17). The fluorescence excitation spectrum of the copper.Ffib fragment (Fig. 7A), obtained by monitoring the fluorescence at 320 nm while scanning the excitation spectrum, showed maximum excitation at 275 nm. Both tyrosine and tryptophan show excitation maxima at this wavelength (18). The fluorescence emission spectrum, obtained by exciting the copper. F.b solution at 270 nm, showed a fluorescence maximum at 319 nm (Fig. 7B), indicating the presence of tryptophan (19). Copper.Fah complex failed to exhibit a characteristic absorption in the EPR spectrum even at a 0.3 mM concentration (based on copper analysis). Prior treatment of the sample with a few crystals of ferricyanide did not result in an EPR signal. A control with 1.0 mM cupric EDTA indicated that 0.05 mM cupric ion would have been detectable. The previous failure to observe an EPR spectrum of the copper. immunoglobulin complex in the presence or absence of ferricyanide (2) could have been explained in terms of spin-pairing of 2 cupric ions. However, the failure to observe a signal with the copper. Fan, which contains but a single copper atom, makes such an explanation less tenable. The amino acid composition of acid-hydrolyzed copper. Fa,, is presented in Table I. from the corresponding spectra of copper. Fah. Moreover, apo-Fak, is very similar to copper. Fab in terms of the large negative Cotton effect at 218 nm (which has been attributed in other proteins to the ,L? structure) and the positive Cotton effect in the 277 to 292 nm region (Fig. 6). Thus, removal of the copper from the cop- per. F,,, complex appears to have little observable effect on structure. Fig. 8 demonstrates that no immunological differences could be detected between apo-Fab and copper. F,h complex or between apo-IgG and copper.IgG complex. The apo-F,,, fragment (Wells 2 and 5) and apo-IgG (Well 4) form a precipitin line of complete identity with the copper.F,k, fragment (Well I) and with the copper.immunoglobulin complex (Wells 3 and 6) when all are allowed to diffuse in agar gel against anticopper. F,b antibodies (center well). This finding supports the conclusion drawn from the similarities of the physical properties of copper. Fab and apo-F,b that removal of copper results Copper-binding Site of a Myeloma Protein a449 isolated chains were shown to be homogeneous by disc gel electrophoresis.
The molecular weights of the light and heavy chains were shown by SDS disc gel electrophoresis to be 28,000 and 56,000, respectively.
The amino acid compositions of the light and heavy chains are shown in Table I. Note that the light chain analyzed for only I S-carboxymethylcysteine residue, whereas the heavy chain analyzed for 3.7 S-carboxymethylcysteine residues (cysteines involved in intrachain disulfide bonds should not have been detected by this procedure). Evidence for the Presence of a Sulfhydryl Group at the Copper-binding Site-Removal of copper from the copper + Fab fragment resulted in exposure of one sulfhydryl group that was very reactive with 4,4'-dithiodipyridine.
Reaction of thiols with this reagent gives the corresponding thiopyridone which has an absorption maximum at 324 nm (21). Fig. 9 shows that copper. Fab reacted to only a small extent with dithiodipyridine, but that apo-F.b reacted to a much larger extent and the absorption generated was proportional to the amount of apo-F,b employed. From the slope of this curve for apo-F.b, using values of 50,000 for the molecular weight of the apo-Fab fragment and 1.4 x lo4 for the molar extinction of the thiopyridone complex, the stoichiometry was calculated to be one reactive sulfhydryl group per apo-F,b fragment. A single determination with reconstituted copper. F,L, (indicated by star) showed that the previously exposed sulfhydryl group was now largely protected as evidenced by the small extent to which the reconstituted complex reacted with the reagent.
in no gross change in tertiary structure. According to this hypothesis, the conformation of the molecule is not dependent on the copper-protein bonds. The failure to observe immunological differences between the apoproteins and their copper complexes also suggests that the copper-binding site is not an antigenic determinant.
Reconstitution of Copper-Fa,, from APO-F~F,~ and Cupric ion-Copper . F,b was reconstituted from apo-F,b by the procedure used to reconstitute copper. immunoglobulin from apoimmunoglobulin (2). APO-F,,, (1.33 mg) in a volume of 1.0 ml was applied to a column (1 x 54 cm) of Sephadex G-25 that had been equilibrated with 0.02 M Tris-HCl, pH 8.0, containing 7.7 PM cupric nitrate. The column was eluted with the cupric nitrate containing Tris buffer. As expected for a successful Hummel and Dreyer run, the level of copper between the trough and the peak was equal to the level before the protein emerged from the column. The peptide peak was eluted with bound copper in a ratio of 1.02 atoms of copper/m01 of Fab. This ratio is consistent with the findings that the isolated and reconstituted copper. immunoglobulin complexes contain 1.8 atoms of copper/mol. Preparation, Purification, and Characterization of Light and Heavy Chains-Treatment of copper. immunoglobulin complex with mercaptoethanol followed by reaction with iodoacetamide yielded the S-carboxamidomethyl derivatives of the light and heavy chain which were subsequently purified by gel filtration in aqueous propionic acid. This procedure has been employed with a number of IgG myeloma proteins for the preparation of derivatives of light and heavy chains in which the intrachain disulfide bonds are still intact (10). The After reaction with dithiodipyridine, apo-F,b was unable to bind copper. No copper was bound to this modified apo-F,t, fragment when it was equilibrated with copper on a Sephadex G-25 column under the exact conditions employed for the successful reconstitution of copper + F,b from unaltered apo- Site of a Myeloma Protein F,,,. Prevention of copper binding by derivatization of the single sulfhydryl group which is exposed by removal of the copper provides evidence that this sulfhydryl is indeed involved in the binding of copper. Alternatively, the inability of the derivatized apo-F,), to bind copper may be a consequence of conformational changes resulting from derivatization, and the failure of copper-F,,,, to react with dithiodipyridine may be explained in terms of a conformational change resulting from binding of copper. However, any conformational changes which do occur as a consequence of copper binding are not detectable by the physical and immunological studies described above.
To ascertain whether the reactive sulfhydryl group detected in the apo-F,h fragment was located on the light chain or on the F,, fragment of the heavy chain, the thiol was labeled by reaction with [14C]iodoacetamide.
Labeling of apo-F,,,, was performed without prior cleavage of disulfide bonds so that only free sulfhydryl groups would be derivatized. Subsequent to the labeling reaction and the removal of labeling reagent, the ['%]carboxamidomethyl-labeled apo-F,,, was reacted with mercaptoethanol and subjected to SDS-polyacrylamide disc gel electrophoresis.
Essentially all of the radioactivity was detected in a polypeptide band with an RF of 0.41. This polypeptide band migrated with the same RF as homogeneous derivatized light chain isolated directly from the copper. immunoglobulin, suggesting that an -SH group of this chain is involved in the copper binding. The fact that, under these conditions, iodoacetamide reacted with apo-F,h but did not react with the histidine residues of the light and heavy chains of the reduced copper.IgG complex (as evidenced by failure to detect carboxymethylhistidine upon amino acid analysis) provides further evidence that it is a sulfhydryl group which is being derivatized in this reaction.
The cysteine involved in copper binding does not appear to be derivatized upon subjecting the intact copper. immunoglobulin to iodoacetamide under conditions where only interchain disulfide bonds are reduced. The light chain derivative prepared in this way contains but a single S-carboxymethylcysteine (see Table I). Under these conditions, other IgG, type immunoglobulins show one light chain carboxymethylcysteine (derived from the cysteine that was involved in disulfide linkage with the heavy chain). It is possible that in the case of the copper. immunoglobulin, the binding of copper to the thiol prevented the thiol from reacting with iodoacetamide and that the additional carboxymethylcysteine would be observed only upon amino acid analysis of carboxamidomethylated light chain derived from apo-IgG or apo-Feb. This is conceivable since the presence of copper prevents the -SH group of F;,), from reacting with dithiodipyridine.
The derivative of the heavy chain obtained by reaction of the intact copper. immunoglobulin with iodoacetamide analyzed for 3.7 carboxymethylcysteine residues. Other IgG, type immunoglobulins have analyzed for 3 such residues (derived from the 2 inter-heavy chain disulfide and the 1 disulfide linkage to the light chain). The detection of the additional 0.7 residue may reflect error in analysis, partial reaction of an intrachain disulfide bond, or the presence of an additional -SH group on the F, fragment of the heavy chain.
Nature of the Copper Complex-The physical properties and the chemical studies give clues as to the nature of this very unusual copper complex. The absence of an EPR signal in the copper. immunoglobulin initially suggested that the molecule contained 2 spin-paired cupric ions or cuprous ions. However, the absence of an EPR signal in the copper. F,k, and the reconstituted copper. F,h complexes, which contain but a single copper atom, makes unlikely the possibility of spinpairing cupric ions. Such a possibility would require that the F,,, fragment be capable of binding 2 copper atoms and that the isolated and reconstituted copper. F,b complexes represent half-saturation of the complex. Although the present data cannot rule out such a possibility, we are working on the hypothesis that each of the two Ffrh fragments of the immunoglobulin binds but a single copper atom. The colorless nature of the immunoglobulin and Fak, copper complexes and the absence of strong absorption bands in their visible and near UV spectra are consistent with the presence of thiol-bound cuprous ions but not compatible with the presence of thiol-bound cupric ions. Proteins and small sulfhy-dry1 compounds containing cupric ion bound to a sulfhydryl group have been shown to exhibit absorption bands in the 350 to 375 nm region (22). Cuprous sulfhydryl complexes, however, generally are either colorless or pale yellow and have no strong absorption band in the near UV region (23-26). The absence of an EPR signal after addition of ferricyanide to copper. immunoglobulin or copper. F,,, indicates that the copper is not oxidized by this agent to cupric ions. This observation would be compatible with the presence of nonoxidizable Cu (1) or Cu(II1) (27, 28).
We do not understand the mechanism by which apo-F,,, binds cupric ions to give a copper complex which is EPR nondetectable.
The reaction of cupric ion with mercaptides to give cuprous ion and disulfide is well documented (25,29,30). The cuprous ion formed is stabilized by formation of a bond with sulfur.
Cd' + 2RS-G CuSR + E( RS), In this reaction, reduction of the cupric to cuprous ion is accomplished with electrons from the mercaptide. In our study of the copper. F,,, fragment, only one sulfhydryl group was detected per F,t, fragment, and each F,b fragment binds 1 copper atom. If copper is being reduced to the cuprous state, the question remains as to the source of electrons.
Conclusions-These studies demonstrate the binding of 1 atom of copper to each F,,, fragment. The inability to detect concanavalin A-binding carbohydrate residues in the copperbinding fragment suggests that the abnormal copper-binding affinity is a property of the unique primary structure of the Fah fragment. Analysis of F,,, and apo-Fab for -SH groups implicates an -SH group in the binding of copper. A labeling experiment indicates that this -SH group is located on the light chain. Other amino acid residues are undoubtedly also involved in the binding of the copper although no evidence for such interactions was obtained from physical studies. The oxidation state of the copper is left unanswered. The ability of apo-F,h to react with a cupric ion to yield an EPR-nondetectable complex makes the copper. F,h complex an excellent and most interesting model for studying EPR-nondetectable copper proteins.