Search

Unlike IgA and IgM, IgG has not yet been shown to form covalent polymers. However in the presence of specific Ag, murine IgG3 has been shown to polymerize through noncovalent interactions. In contrast to the noncovalent oligomers found with murine IgG3, we have detected covalent dimers in three different recombinant human IgG2 Abs produced in myeloma cells. Both IgG2,kappa and IgG2,lambda can form dimers. In addition, analysis of pooled human gamma globulin and several normal sera revealed the presence of IgG2 dimers. The IgG2 dimers are in contrast to the noncovalent IgG dimers found in pooled sera of multiple donors resulting from idiotype/anti-idiotype (Id/anti-Id) interactions. Cyanogen bromide cleavage analysis suggests that one or more Cys residues in the gamma 2 hinge are involved in dimer assembly. The potential role of IgG2 dimers in immunity against carbohydrate Ags is discussed.

A considerable part of the Ab repertoire is given over to polyreactive Abs capable of interacting with multiple antigenic species. Neither the function of these Abs nor the molecular basis for their activity is known. To address the latter problem, we have compared the amino acid sequences of a large panel (n = 70) of polyreactive human monoclonal Fab fragments and conducted a series of engineering experiments on a prototype polyreactive Fab. The Fab fragments were retrieved from combinatorial IgG libraries prepared from the bone marrow of long term asymptomatic HIV-1 seropositive donors. The general features displayed by the panel of IgG polyreactive Abs include 1) skewed VH family usage with a predominance of VH1 and VH4 clones and a paucity of the normally prevalent VH3 family; 2) use of a variety of different VH germ-line genes within the context of the family usage and no restriction in D or JH gene usage; 3) skewed VL gene usage: 75% of Fabs used one of two germ lines; and 4) extensive somatic modification of both heavy and light chains. The importance of the heavy chain, in particular the heavy chain CDR3 (HCDR3), in dictating the polyreactive phenotype was demonstrated for the prototype Fab by chain shuffling and CDR transplantation experiments. in addition, and most strikingly, a constrained peptide based on the HCDR3 sequence was shown to be polyreactive and to inhibit binding of the parent Ab to a panel of Ags. A role for conformational flexibility in polyreactivity was suggested by a marked temperature dependence of Ab recognition of Ag. One Ab was shown to be polyreactive at 37 degrees C, but was apparently monoreactive at 4 degrees C. We hypothesize that Ab polyreactivity is associated with conformationally flexible HCDR3 regions in the context of certain favorable framework configurations.

The genes encoding Ig chains in elasmobranchs are arranged in clusters or cassettes, an organizational pattern dramatically different from the mammalian translocon gene arrangement. Cluster gene arrangements, which have now been found in non-elasmobranchs as well, pose interesting dilemmas for understanding the mechanisms of Ig gene expression and regulation in terms of allelic exclusion and clonal selection. We have sequenced five lambda genomic clones encoding complete sandbar shark (Carcharhinus plumbeus) lambda L chain gene loci. While the coding regions among all five clones are highly homologous, the noncoding regions have significant differences that allowed us to identify two types of lambda L chain clusters. The noncoding regions are < 60% identical between groups, while the three clones belonging to the first group share > 95% identity in their noncoding regions. The second group is more diverse and may be comprised of several related subgroups. The two clones in this group share approximately 85% identity in the noncoding regions. Variations in the promoter region, including octamer and TATA box orientation and position, are identified between the two groups and may have implications for the molecular regulation of Ab production. Our results show the sandbar shark lambda L chain family to be a complex and diverse system.

Monoreactive high affinity pathologic autoantibodies were supposed previously to derive through somatic mutation from polyreactive low affinity autoantibodies that are encoded by a small set of unmutated V region genes in fetal and neonatal B cells. However, recent data exploring the physiologically expressed Ab repertoire and the importance of the stochastically generated heavy chain CDR3 (H-CDR3) in autoreactivity suggest that this scheme is incomplete. Here we analyzed via gene-swapping experiments and site-directed mutagenesis the relative contributions of the mutations in the light chain variable region (VL) and the heavy chain variable region (VH) domains and of the H-CDR3 in the autoreactivity of two IgM rheumatoid factors (RF), one a polyreactive low affinity Ab, the other a monoreactive high affinity Ab. These two RFs derived from the same V kappa III (humkv325) and VH1 (51p1) genes, but differed from each other by a few mutations and by the structure of the H-CDR3. The analysis of the reactivity patterns of different combinations of wild-type and in vitro engineered hybrid gene products clearly demonstrates the main influence of the H-CDR3 in the autoAb activity profiles. The results directly demonstrate the previously proposed hypothesis, namely, that the H-CDR3 plays a critical role in distinguishing poly- from monospecific RF. However, the data also indicate that self polyreactivity is a very fragile property and is dependent upon the primary structure of the VH segment.

Three cross-reactive idiotopes(Id), termed IdX, IdI-1, and Id5, that are present on free L chains from murine anti-group A streptococcal carbohydrate antibodies have been mapped; these Id distinguish between products of three homologous V kappa genes. For each determinant, sequence analysis of anti-streptococcal group A carbohydrate antibody V domains yielded small numbers of amino acids invariably associated with Id expression. Flow micro-fluorimetry was used to isolate three IdI-1- spontaneous mutants of the IdI-1+ hybridoma GAC 39; all had single amino acid changes in the L chain at position 60 and 77, all retained other Id, and all bound group A carbohydrate. Computer modeling was used to examine spatial relationships between Id. A number of the conserved Id5 and IdX residues cluster in the L chain framework region 1 around the first back loop connecting strands of the beta pleated sheets, and overlap at residue 15 (Id5, proline; IdX, leucine). This overlap accords with the mutually exclusive expression of Id5 and IdX. The IdI-1 loss variants have mutations of residues 60 or 77 on adjacent back loops, approximately 7.5 and 14 A from residue 15. Competitive inhibition of anti-IdX and anti-IdI-1 binding to antibodies expressing both Id can be attributed to steric hindrance. The framework back loops may be favored sites for cross-reactive Id expressed by products of a single V region gene. IdI-3a, an individual Id not associated with use of a particular gene segment, has been localized in part to residue 31 (hypervariable region 1) of the H chain.

We report on the detection of the same idiotype (Id62) on the heavy and light chains of a mouse monoclonal autoantibody to thyroglobulin. This result was obtained by a simple, sensitive, and reproducible technique, namely immunoblot on nitrocellulose paper of electrophoretically separated heavy and light chains by using a panel of 125I-labeled anti-idiotype probes. These included polyclonal heterologous and syngeneic affinity-purified antibodies and a syngeneic monoclonal antibody. The immunoblot findings were confirmed through the solid-phase radioimmunoassay binding of the anti-idiotype probes to isolated H and L polypeptide chains prepared by conventional chromatography methods. Because anti-idiotype probes define a putative regulatory idiotype of the BALB/c autoimmune response to thyroglobulin, it is suggested that the simultaneous presence of the same idiotype on H and L chains may confer to antibody regulatory properties within the idiotypic network. The significance of the existence of the same idiotype on H and L chains and the implication of this finding for the regulation of the immune response to self antigens are discussed.

A large fraction of the anti-p-azophenylarsonate antibodies of strain A/J mice share a major cross-reactive idiotype (IdCR). Structural analysis of monoclonal antibodies expressing this idiotype (IdCR+) indicates that a particular combination of variable region gene segments (Vk, Jk, VH, D, and JH) encodes the variable regions of the light and heavy chains of these IdCR+ antibodies. With the use of serologic methods, hybridoma cell lines have been isolated that produce monoclonal antibodies lacking IdCR determinants (IdCR-), but that are derived from most of the same combination of variable region gene segments that encode IdCR+ monoclonal antibodies. Structural analysis of these IdCR- monoclonal antibodies demonstrates that they are very homologous to each other and to IdCR+ monoclonal antibodies with respect to VH and VL sequences, but are markedly different from IdCR+ monoclonal antibodies in their utilization of D region segments. Comparisons of antigen avidity of these IdCR+ and IdCR- antibodies indicates that conservation of D region structure is not crucial for effective antigen binding. These results indicate the importance of the D region in idiotypy in the IdCR system and demonstrate the variation permitted in D region structure while maintaining antigen recognition.

kappa-Myeloma antigen (KMA) was immunoprecipitated from lactoperoxidase-radioiodinated HMy2 lymphoblastoid cells by using monoclonal antibody K-1-21 and was analyzed by SDS-PAGE. Under reducing conditions, two major subunits of Mr approximately 26,000 and Mr approximately 42,000, and minor components of Mr approximately 28,000, 31,000, and 36,000 were observed. The Mr approximately 26,000 subunit was identical to kappa-light chains from HMy2 surface IgG in apparent m.w., isoelectric point, and staphylococcal V-8 protease peptide map, but was not precipitated in association with Ig heavy chain. The Mr approximately 42,000 component was homologous to rabbit skeletal muscle actin by peptide mapping with staphylococcal V-8 protease. The cell surface origin of the immunoprecipitated antigen was confirmed by demonstrating lactoperoxidase dependence of iodination and complete removal from the cell surface after pronase treatment of viable cells. Thus, cell surface expression of KMA is the result of membrane association of non-heavy chain-linked kappa-light chains, possibly in noncovalent association with actin.

The immunoglobulins of the bullfrog Rana catesbeiana are unusual in that, in all classes, the light chains are not disulfide bonded to heavy chains or to other light chains. Moreover, the light chains contain six, rather than the usual five, residues of half-cystine. As none of these half-cystines is in the sulfhydryl form or is alkylated after mild reduction, we suggested that the light chains probably contain three intrachain disulfide bridges. We have now carried out experiments to confirm the existence of an extra intrachain disulfide bridge in Rana catesbeiana light chains and to determine its location. Disulfide bridge assignments were based on 1) isolation and sequence analysis of S-(carboxymethyl)cysteine-containing peptides and 2) isolation, from unreduced light chains, of peptides containing a disulfide bridge. Half-cystine residues were found at positions 134 and 194, and these were shown to be joined in the conserved intradomain disulfide bridge. In addition, we found that a residue of half-cystine, located at the third position from the carboxy-terminus, forms a disulfide bridge with a half-cystine at position 119, near the amino-terminus of the domain, the latter residue replacing a proline that has been found at this position in all other light chains. An intrachain disulfide bridge has not been found at this location in any other light chain.

As part of understanding molecular function in structural terms, we have been attempting to map the idiotypic topography of specific anti-arsonate (Ars) antibodies. A panel of anti-Ars hybridomas of which the complete primary sequences are known were used. These molecules show a varied reactivity profile with a panel of monoclonal anti-idiotypic antibodies. By judicious chain recombination experiments and chemical modifications that altered this reactivity profile, we were able to identify particular amino acid residues or discreet regions of anti-Ars antibodies as having crucial roles in the expression of idiotypic determinants. Idiotopes were mapped to the heavy chain second hypervariable region and D segment, and to the light chain first and third hypervariable regions.

The murine VH genetic map consists of two different subloci of VH genes. Seven idiotypes have been allocated to these two different subloci, one of which contains genes coding for the A5A idiotype, while the second contains the genes encoding for six other idiotypic markers. Previous studies from this laboratory have shown that anti-Ars antibodies are best classified in the murine VHII subgroup. Studies from other laboratories indicate that the J558 idiotype could be classified in a similar fashion. The sequence of the T15 myeloma protein VH region is most characteristic of the murine VHIII subgroup. These studies were undertaken in order to study the VH subgroup of two additional murine VH markers. Since one of these markers has been allocated to a separate sublocus, comparisons can be made for the first time between the "framework" residues (VH subgroup) and the murine genetic map. The data indicate that within a single presently defined sublocus proteins exist which contain remarkably different "framework" residues.

Sjogren's syndrome presents a wide spectrum of disease manifestations ranging from benign to malignant lymphoproliferation. Sera from 21 patients with primary Sjogren's syndrome, 63 patients with other autoimmune rheumatic diseases, and 140 normal controls were studied by using high-resolution gel electrophoresis combined with immunofixation and specific absorption studies. Two-thirds of the sera from Sjogren's syndrome patients contained free monoclonal lambda light chains. In addition, one patient with Sjogren's syndrome and pseudolymphoma had a circulating monoclonal IgM-lambda immunoglobulin. In contrast, only 16% of the patients with other autoimmune diseases and 5% of normals had monoclonal bands; none of the other patients studied exhibited monoclonal-free lambda light chains in their serum. Our findings suggest that the monoclonal process in Sjogren's syndrome starts early in the disease process in a subset of B cells.

The cross-reactive idiotype (CRI) associated with the BALB/c antibody response against the p-azophenylarsonate (Ar) hapten was studied by analyzing hybridoma anti-Ar derived from BALB/c mice. The BALB/c CRI (CRIc) was previously thought to be a single idiotype as defined by rabbit anti-idiotype. CRIc has been resolved into at least two separate families of anti-Ar antibodies that are unrelated idiotypically to each other. Analysis of the 5AF6 family revealed considerable idiotypic heterogeneity, including a number of public idiotopes that appeared to be primarily associated with combinatorial idiotopes (requiring H + L chains) or L chain-specific idiotopes.

Patients with B cell neoplasms frequently have low levels of tumor-related light chains in their urine; these light chains can be isolated with the use of relatively simple methods and then used to raise antibodies to the idiotypic determinants. In this study, anti-light chain idiotypes were raised against monoclonal light chains from the urine of four patients with chronic lymphocytic leukemia. The antibodies reacted specifically with the tumor cells of the homologous patient, assessed by immunofluorescence, and can therefore be used for tumor cell detection. In one case for which serum idiotypic IgM was available, the anti-light chain idiotype was shown to bind whole idiotypic IgM, and such binding could be inhibited by idiotypic IgM or idiotypic light chains, which demonstrates recognition of similar antigenic determinants. The binding of antibody to tumor cells was also totally inhibited by idiotypic IgM. The analysis of separated sera from the four patients for free light chains demonstrated only low levels (3.0 to 8.6 micrograms/ml of serum with a mean of 5.8), which suggests that light chain is rapidly cleared and therefore does not present a major barrier to antibody attack. It should be feasible to use such antibodies for both analysis and therapy of B cell neoplasms.

The amino terminal (positions 0 to 20) amino acid sequences of 22 rabbit antibody light chains are reported and compared with 44 amino terminal sequences previously described in the literature. Rabbit k light chains demonstrate three different amino terminal chain lengths and considerable variability at the amino terminal end. In relating amino terminal sequence to b locus allotype, allotype-specific residues (amino acid residues unique to all light chains having a given allotype) were not identified, although certain residue alternatives were unique to some of the b5 and b9 light chains. There was no correlation of antibody specificity with amino terminal sequence. In the most striking example, identical amino terminal sequences were associated with two different carbohydrate specificities (type VIII pneumococcal polysaccharide and streptococcal group A carbohydrate) and an aminophenyltrimethylammonium specificity. An obligatory association of amino terminal sequence with complete variable region framework sequences was not observed. Thus, it is not always possible to predict the framework homology among a given pair of light chains by examination of their amino terminal sequences. The limited diversity seen among the variable regions of murine myelomas that have the same antigen-binding properties is not found in the light chains of specifically elicited rabbit antibodies.

Amino terminal sequence analyses were performed on the H and L chains of the idiotypically related IgMlambda and IgGkappa paraproteins isolated from the sera of a patient, Br. The N-terminal 41 residues of the Br k-chain belonged to the Vkiii subgroup, and L chains derived from BrIgMlambda revealed a blocked N-terminus characteristics of lambda-chains. Comparative sequence analysis of the Br mu- and gamma-chains indicated that, although both possessed an unblocked N-terminal glutamic acid, the respective VH regions belonged to seperate subgroups. The N-terminal 27 residues of the Br mu-chain reflected a typical VHiii subgroup sequence. The Br gamma-chain sequence demonstrated a VHI pattern with an unblocked N-terminus. These structural data stand in contrast to previously reported serologic evidence, which indicated that the BrIgMlambda and BrIgGkappa proteins possessed highly similar Vh-associated idiotypic determinants. Final interpretation of these findings of similar Vh idiotypic determinants expressed by H chains belonging to seperate Vh subgroups must await complete sequence analysis of hypervariable segments of the Br gamma- and mu-chains.

Four homogeneous antibodies to type VIII pneumococcal polysaccharide (S8) were isolated from the serum of a single rabbit (3322) by affinity chromatography on an S8 immunoadsoebent by utilizing gradient elution with cellobiose and NaCl. The binding properties of these antibodies were determined by a radioimmunoassay with 125I-bovine gamma-globulin-S8. Cellobiose (a disaccharide unit of S8) was the immunodominant group of each of the four antibodies, but each antibody bound to this disaccharide with different relative affinities. The amino acid sequences (positions 0-40) of three of the four antibody light chains were each different both in framework and first hypervariable region sequences. The fourth antibody light chain has a blocked amino terminus. These findings indicate that antibodies elicited by a relatively simple antigen and examined at one time during the course of immunization in a single rabbit may exhibit common specificities for an oligosaccharide determinant, yet have different binding affinities for that determinant as well as different primary structures in the complementarity (hypervariable) regions and framework regions.

A rabbit anti-idiotypic antiserum was raised against a monoclonal human IgM kappa(Me) in order to analyze the possible modulation of idiotypic expression by Fab constant domains. IgM(Me) fragments, subunits, and domains were prepared by chemical and enzymatic cleavages. All molecular species were shown to have a well-defined secondary and tertiary structure by circular dichroism. Full recombination between domains and subunits was ascertained by difference spectroscopy. The expression of the idiotype on native and recombined fragments, domains, and subunits was quantitated in a competitive enzyme-linked immunosorbent assay (ELISA). Reduced and alkylated Fab, isolated H and L chains, purified Fv(Me), intact VH and VL domains and H-L, VH-L, VL-H, and VH-VL recombinants were compared on a molar basis to native Fab(Me) for idiotypic expression. VH-specific determinants were found, whereas the L chains were virtually devoid of idiotypic activity. Both the peptic FV(Me) fragment, which is composed of intact VH and VL domains, and the recombined VH-VL heterodimer were found to be fourfold less active for idiotype expression than native Fab(Me). However, full inhibition was achieved at high molar concentrations, suggesting that all the idiotopes present on Fab(Me) were expressed on FV(Me) but with a reduced antigenicity. Comparison of VH-L and VL-H hybrid molecules revealed that the presence of the C mu 1 domain was sufficient to restore full idiotypic expression as compared with native Fab(Me). These data support the hypothesis that the first constant domain of the mu heavy chain alters the quaternary interaction between the variable domains, and therefore modulates the expression of the idiotype through longitudinal interactions that are not affected by reduction of the inter-H-L chain disulfide bond.

The L chains of rabbit antibodies directed against group a allotypes exhibited in many instances a high degree of homogeneity as measured by L chain banding patterns in alkaline urea polyacrylamide gels. Amino terminal sequence analyses were carried out on six L chain preparations from antibodies isolated from individual antisera or from pools of antisera. The same major amino terminal sequence, Ala-Val-Val-Met, was observed for each of these preparations indication that anti-allotype antibodies preferentially select L chains from a single subgroup. The antiallotype anitbodies were isolated from antisera by elution from IgG immunoadsorbent columns in yields ranging from 0.3 to 1 mg/ml antibody. The specificity of the isolated antibodies was demonstrated by radioimmune assays. Certain fractions were contaminated with a protein that had properties similar to rabbit serum albumin. This contamination was minimized by preadsorption of the antisera. The antibodies were primarily of the IgG class as shown by immunoelectrophoresis and by m.w. of the H and L chains on SDS gels.

The VK1GAC light chain represents the dominant V kappa structure employed in the antibody response of A/J mice to streptococcal group A carbohydrate ( GAC ). Two anti-idiotypic antisera, anti- Id5 and anti- Id20 , with specificity for the VK1GAC light chain were used to examine anti- GAC antibody responses in a series of inbred mouse strains that differ at the heavy chain constant region ( IgCH ) allotype locus. Both idiotypes were expressed in normal and immune sera from mice of most IgCH allotypes, except IgCHb (C57BL/6J) and IgCHf (CE/J). C57BL/6J mice expressed Id5 , but not Id20 , whereas CE/J mice did not express either idiotype. Testing of recombinant inbred strains between BALB/c and C57BL/6 indicated that the pattern of idiotype expression did not correlate with IgCH allotype. The C X B recombinants expressed all three idiotype patterns that were observed in the panel of inbred strains. Testing of allotype congenic mice between BALB/c and C57BL/6 showed that CB.20 and BC.8 mice were Id20 -, whereas BAB-14 mice were Id20 +, indicating that both VH and background (V kappa or regulatory) loci must be derived from BALB/c to obtain Id20 expression. The difference in the frequency of idiotype expression observed between BALB/c and BAB-14 mice indicates that the IgCH locus may exert a quantitative influence on the expression of this light chain. To examine the Id20 -, Id5 + antibodies of C57BL/6 mice, anti- GAC hybridomas were prepared. Of 16 C57BL/6-derived anti- GAC monoclonal antibodies, six were reactive with anti- Id5 and not with anti- Id20 . Isoelectric focusing of the purified kappa light chains from three of these antibodies revealed two distinct spectrotypes that co-migrated with the two known VK1GAC spectrotypes observed with A/J anti- GAC light chains. Idiotypic analysis of in vitro recombinants between the heavy and light chains of A/J and C57BL/6 monoclonal antibodies demonstrated that the C57BL/6 light chains were idiotypically similar to A/J light chains when they were free in solution or paired with A/J heavy chains. These results demonstrate that C57BL/6 mice can express a light chain that is very similar, if not identical, to the VK1GAC light chain, although the light chain is expressed in lower frequency and is paired with a distinct VH structure, which can mask expression of one of the VK1GAC idiotypes. These effects on V kappa expression map to at least three genetic loci: VH, CH, and an unlinked locus.

The amino terminal sequences of the kappa light chains from SAMM 368 IgG2b and IgA have been determined. The two chains belong to different kappa subgroups as indicated by 10 amino acid differences among the first 23 residues. An additional three residue differences were demonstrated in the sequence including the first complementarity region. These results indicate that single cells of SAMM 368 produce two unique kappa-chains.

The complete amino acid sequence of the human monoclonal lambda VI light chain Bence Jones protein THO was determined. We have found it to have remarkable similarities to the previously sequenced lambda VI Bence Jones protein SUT. Immunochemical analyses demonstrated that both lambda VI chains belong to a V lambda VI sub-subgroup. The 98-residue V gene-encoded segments of proteins THO and SUT are closely homologous and are distinguished from other lambda VI chains by a one-residue deletion at the V-J recombination site. Proteins THO and SUT have identical 13-residue J segments and therefore are encoded by the same J lambda gene. Further, both proteins have identical 105-residue C regions that by sequence represent products of the C lambda 3 (Kern-, Oz+) gene. The primary structure and serologic properties of proteins THO and SUT imply at the protein level of association between certain types of V lambda, J lambda, and C lambda segments.

It has been suggested that light chains (LC) are necessary for the surface expression of mu heavy chains. Fluorescent antibody screening of 42 human lymphoblastoid cell lines transformed in our laboratory, however, disclosed four lines that expressed surface mu-chains without LC. The biosynthesis, glycosylation, and turnover of mu-chains in these cell lines was compared to mu-chain production in cell lines synthesizing both heavy chains and LC. LC production could not be detected in the mu +LC- cell lines by either surface or biosynthetic labeling. The mu-chains expressed on the surface of the LC- cells appeared as disulfide-linked dimers and migrated slightly faster on SDS-polyacrylamide gels (70 Kd) than did mu-chains from IgM monomers (H2 L2) (78 Kd) after reduction. Biosynthetic labeling in the presence of tunicamycin demonstrated that the smaller size of free mu heavy chains was due to incomplete glycosylation of these molecules and not to amino acid deletions. The mu-chains produced by mu + LC- cells lines were degraded faster than mu-chains from LC+ cell lines, but their rate of transit to the cell surface was identical in both cell types. Thus, although LC are necessary for the formation of an intact antigen-binding site, they are not involved in the synthesis or expression of membrane mu-chains.

K-1-21 is a murine monoclonal antibody that reacts with human kappa-light chains in free form but not when they are associated with immunoglobulin heavy chains. K-1-21 was unexpectedly shown to bind to a determinant, STA (Sezary T cell antigen), detected by immunofluorescence in the cytoplasm but not on the surface of Sezary T cells isolated from peripheral blood (4/4 cases) and in Sezary T cells from lymph node and bone marrow (one patient). STA was detected in F2/F7, CCRF-CEM, Molt-4, and CCRF-HSB (four human T ALL cell lines), in JURKAT (a human T cell leukemia line), and in MLA144 (a Gibbon T cell lymphoma line). It also occurred in Leu-3a+ antigen-specific T cell clones (6/6 tested). Moreover, although STA was absent from freshly isolated normal T cells, its expression could be evoked in E+ cells from peripheral blood by in vitro culture with phytohemagglutinin. Thus, STA appears to be a cytoplasmic marker for activated T cells. Cytoplasmic inhibition immunofluorescence studies indicated that K-1-21 binding to STA in Sezary cells or T cell lines was inhibited by preincubation of the K-1-21 antibody with purified kappa-Bence Jones protein. STA from radiolabeled MLA144 cell lysates was immunoprecipitated by K-1-21 and was identified on polyacrylamide gel electrophoresis under reducing conditions as a protein of m.w. 57,000. Additional experiments are underway to define the molecular basis of the interesting cross-reactivity between a determinant in T cells and the K-1-21 reactive epitope on free kappa-light chains.

To assess the heterogeneity of immunoglobulins involved in various skin diseases, direct and indirect immunofluorescence studies of skin biopsies and sera, respectively, for kappa and lambda light chains, were performed. The anti-basement membrane zone (anti-BMZ) antibodies of patients with bullous pemphigoid showed a predominance of kappa light chains, and patients with linear IgA bullous dermatosis showed a predominance of one light chain that was sometimes kappa and sometimes lambda. The bullous pemphigoid autoantibodies were then studied for IgG subclass distribution; a predominance of IgG4 was found. Although other explanations are possible, the light chain restriction in bullous pemphigoid most likely reflects heavy chain restriction and preferential association of heavy and light chain isotypes. The basis of the heavy chain restriction is not apparent. The light chain restriction in linear IgA bullous dermatosis may represent a restricted idiotypic repertoire.

125I-membrane IgM, 125I-membrane IgD-like molecules, and their constituent chains from iodinated murine splenocytes were characterized by sodium dodecyl sulfate polyacrylamide gel electrophoresis. The apparent m.w. of the heavy chains decreased as the acrylamide concentration was raised. The membrane mu-chain had a slower mobility than did mu-chain from secreted IgM. Unreduced IgM and IgD-like molecules had mobilities consistent with an H2L2 structure. Intact IgD-like molecules were replaced after overnight dialysis by molecules with the properties of HL. Unreduced surface IgM had a slower mobility than that of monomeric IgM obtained by partial reduction of secreted IgM.

The serum antibodies from humans booster-immunized with diphtheria toxoid were analyzed by isoelectric focusing. To restrict the antibody response, we visualized those antibodies that reacted to the Fragment A moiety of diphtheria toxin. Of the 100 normal human donors examined, 20 were estimated to be nonresponders to Fragment A. Sixty-three of the 80 donors who responded to Fragment A had restricted IgG-anti-Fragment A spectrotypes that consisted of three to eight distinct bands. Six series of repeat or shared spectrotypes were delineated among our donor population with repeat frequencies ranging from 0.025 to 0.312. We determined that the repeat spectrotypes were IgG1-kappa anti-Fragment A antibodies. Three percent of our donors had complex IgG-anti Fragment A spectrotypes consisting of greater than 10 bands. In some instances, the complex spectrotypes were composed of simple spectrotypes found among other donors. The complex IgG-anti Fragment A spectrotypes of these donors consisted of both IgG1 and IgG4-anti-Fragment A antibodies. From a statistical analysis, we calculate that only a limited number of IgG-anti-Fragment A spectrotypes are expressed in vivo after booster immunization. Our analysis suggests two groups of IgG-anti-Fragment A spectrotypes exist in our donor population. One IgG-anti-Fragment A spectrotype group is expressed randomly, and a second group is preferentially expressed among individuals in our donor population.

The complete amino acid sequence of the variable regions of light chains derived from anti-p-azophenylarsonate antibodies from A/J mice bearing a cross-reactive idiotype is reported. At least two and probably more than three distinct light chains are associated with this idiotypically characterized antibody. The antibodies have several differences in their "framework" structures but evidence is presented indicating that all three light chain hypervariable regions have a homogeneous sequence. The data are discussed in relation to the various theories of antibody diversity. In addition, the findings support the view that hypervariable regions, idiotypic determinants, and the antibody-combining site involve, to a large extent, the same molecular structures.

Many monoclonal antibodies that react with the lacto-N-fucopentaose III (LNF III) antigenic determinant, Gal beta 1-4(Fuc alpha 1-3)GlcNAc beta 1-3Gal beta 1-4Glc, have been described recently. The terminal trisaccharide of this determinant, fucosyllactosamine, is present on glycolipids and glycoproteins and on the surface of granulocytes, monocytes, and other cells. To study the structural and genetic diversity of these antibodies, syngeneic anti-idiotypic monoclonal antibodies were produced in BALB/c mice against PMN 6, a monoclonal antibody directed against this sequence. Anti-idiotypic antibodies 6B1 and 6C4 reacted with 50% of a panel of 20 anti-LNF III monoclonal antibodies, whereas 6A3 reacted strongly only with PMN 6. This indicates that the determinants recognized by 6C4 and 6B1 represent major cross-reactive idiotopes of this family of antibodies. The binding of idiotypic antibodies to a glycolipid bearing this antigenic determinant was completely inhibited by the three anti-idiotypic antibodies, 6A3, 6B1, and 6C4. The idiotopes could be demonstrated on the heavy chain of the monoclonal antibodies by an antibody transfer technique when mild reducing conditions were employed, but a high concentration of reducing agent destroyed the idiotypic determinants. This suggests that the anti-idiotypic antibodies recognize conformational structures expressed on the heavy chain molecules. The binding of 18 monoclonal antibodies to two glycolipid antigens and to a fucosyllactosamine-bovine serum albumin conjugate was compared. Antibodies that possessed the 6C4 cross-reactive idiotope bound to fucosyllactosamine-bovine serum albumin more weakly than idiotype-negative antibodies (p = 0.001). This suggests that the 6C4-positive antibodies might represent germline structures.

Immunoglobulin kappa-chain mRNA was hybridized with DNA in order to assess the kappa-gene frequency. Kappa-mRNA was purified from membrane-bound ribosomes of mouse myeloma MOPC-41 by poly (U) chromatography and isolation of a 13S RNA by successive sucrose density gradient centrifugations. The RNA coded for kappa-chain precursor molecules in cell-free protein synthesis and essentially no other proteins. MOPC-41 kappa-mRNA hybridized with MOPC-41, MPC-11, and Krebs DNA with the same kinetics: the majority of the hybrids was formed with rare or unique DNA sequences (Cot/2 450 to 900), a small portion with highly repetitive sequences (Cot/2 5--6). The slow hybrids were well matched and the rapid hybrids were mismatched by about 4%, regardless of the DNA used. It was further investigated whether the rapid hybrids contained translatable kappa-mRNA or were due to impurities in the RNA preparations. Kappa-mRNA and globin-mRNA (as an internal standard for a unique transcript) were hybridized with DNA to Cot 20 or 48, the hybridized and unhybridized RNA were isolated by hydroxyopatite-urea chromatography and, after removal of the DNA, translated in a cell-free system. The cell-free products were analyzed by SDS-polyacrylamide gel electrophoresis and immunoprecipitation. It was found that approximately equal quantities of translatable kappa- and globin-mRNA were hybridized maximally 1.7%). The results do not support the hypothesis that kappa-mRNA is a transcript of both repetitive and unique DNA sequences.

Cultures of peripheral blood lymphocytes (PBL) in which specific hemolytic plaque-forming cells (HcPFC) had been induced were labeled with 14C-amino acids. Antigen-specific products in the culture supernatants were characterized by using indirect immune precipitation in conjunction with specific immunoabsorbents and/or gel filtration followed by SDS-polyacrylamide gel electrophoresis. After 5 days of culture with antigen (sheep red blood cells or ovalbumin) newly synthesized IgM and specific IgM antibody were demonstrated in culture supernatants from normal donors and from four out of five patients with congenital agammaglobulinemia (cAgamma). Secreted products bound specifically to antigen and pretreatment of labeled supernatants with anti-mu and anti-L chain antisera, but not with anti-gamma antiserum, prevented binding. Typical mu- and L chains constituted only a proportion of the anigen-binding peptides recognized by the anti-mu reagents. Induction of IgM antibody synthesis was dependent on the presence of antigen and was correlated with the generation of HcPFC. No major differences between the antigen-induced products of cAgamma and normal PBL were observed. These findings suggest that in the absence of terminal B cell differentiation in vivo, certain patients with cAgamma possess precursor cells that can respond to antigen in vitro with the synthesis of specific humoral products, including IgM antibody.

The complete amino acid sequence of five light chain variable (V) regions of human monoclonal IgM kappa rheumatoid factors (RF) was determined, and their cross-reactive idiotypes (CRI) were characterized with antibodies induced by immunization with synthetic peptides PSL2 and PSL3, corresponding to the second and third complementarity-determining regions (CDR) of the SIE light chain. Together with two additional RF studied previously, all seven RF belong to the V kappa IIIb sub-subgroup. The region encoded by the V kappa gene segment (positions 1 to 95) in all seven proteins was virtually identical in primary structure, whereas the sequence from positions 96 to 108 defined the usage of the J kappa 1 gene in three proteins and the J kappa 2 gene in four of them. Position 96 contributed by the recombination of the V kappa and J kappa gene segments showed the presence of four different amino acid residues. Both anti-PSL2 and anti-PSL3 bind efficiently to all separated L chains when analyzed by the Western blot technique, and the binding was inhibited specifically by the corresponding peptides. The results reveal that the majority of human IgM-RF light chains are derived from a single germ line V kappa gene or a family of closely related V kappa III germ line genes, and express two "primary structure-dependent" CRI, which are largely dependent on the amino acid sequence of the second and third light chain CDR.

Exposure of Fc fragments derived from human IgG1 myeloma proteins to acid pH rendered the region between the Cgamma2 and Cgamma3 domains transiently susceptible to cleavage by trypsin upon return to neutral pH. Trypsin covalently linked to Sepharose was used and two fragments derived from the Cgamma2 region and one from the Cgamma3 region were purified by column chromatography. On the basis of amino acid analysis, primary sequency data, antigenic properties, and m.w., one of the Cgamma2 fragments was shown to consist of two polypeptide chains of unequal mass joined by the inter-heavy chain disulfide bonds. The larger chain corresponded to a stretch of gamma-chain between Thr 223 and Lys 338 (Eu numbering) and the shorter chain to the section between Thr 223 and Lys 248. The other Cgamma2-fragment was a disulfide-linked dimer of the Thr 223 to Lys 338 sections of the paired gamma-chains. When this latter fragment was reduced under mild conditions it dissociated into monomers indicating that there was little or no noncovalent interactions between the Cgamma2 domains. The Cgamma3-fragment was shown to be a noncovalent dimer composed of the Glu 345 to Lys 349 sections of the two gamma-chains although some heterogeneity was apparent at the amino-terminus. Circular dichroism was used to probe the conformational relationships between the isolated domains and the parent Fc. The spectral properties of Fc could not be fully accounted for on the basis of the spectra observed for the isolated domains which suggested that inter-domain interaction might be significant in Fc.

We have employed five spectrotypically distinct monoclonal anti-variable region antibodies in the definition and characterization of a set of idiotopes expressed on murine monoclonal antibodies specific for streptococcal group A carbohydrate (GAC). By evaluating which of a panel of monoclonal anti-GAC antibodies were bound by the various anti-idiotopes, we observed four distinct reactivity profiles for the five anti-idiotopes ranging from highly restricted (binding of the homologous anti-GAC monoclonal antibody only) to broadly cross-reactive (binding of 18 of the 38 IgG3 anti-GAC antibodies). With N-acetyl-D-glucosamine and soluble GAC used as haptens, this spectrum of reactivity profiles was paralleled by a gradient of susceptibility to hapten inhibition of anti-idiotope binding to idiotope. The degree of cross-reactivity exhibited by a given anti-idiotope was found to be inversely related to its susceptibility to hapten inhibition. The topographic relationships among the idiotopes, defined by the results of competitive binding assays, were suggestive of a linear idiotope map spanning the variable region from the antigen-binding site to the vicinity of the constant region. Additional data from competitive inhibition assays with isolated and recombined H and L chains from a prototype monoclonal anti-GAC antibody (HGAC 39), and from isoelectric focusing of whole or reduced and alkylated HGAC 39, suggested that one of the idiotopes was located, at least primarily, on the VL domain.

Anti-idiotypic immunization triggers the production of antibodies that are structurally related to the idiotype. We have shown that the heavy chain variable regions of antibodies produced after anti-ABPC48 (A48) anti-idiotypic immunization of BALB/c mice are homologous to that of A48, except for the third hypervariable region. We present here partial light chain sequences of A48 and of antibodies induced by anti-idiotypic immunization. Nearly perfect homology is found, suggesting that these chains are the products of genes derived from a unique VK germ-line gene. These observations indicate that the H and L hypervariable regions contribute to define the structure of A48 idiotopes. Remarkably, the VK sequence we identify is the same as that described for anti-arsonate and anti-oxazolone antibodies. We discuss the relative importance of particular amino acids for idiotype expression and antigen-binding activity.

Light chains isolated from normal immunoglobulin give rise to a finite number of discrete bands when analyzed by isoelectric focusing in polyacrylamide gel. The positions and relative intensities of the bands were identical in light chains isolated from different individuals of a species, but clear differences were evident when light chains of several mammalian species were compared.

We have dissected the human influenza-specific B cell repertoire by performing Epstein-Barr virus (EBV) limiting dilution analysis of lymphocytes obtained from donors before and after immunization with a commercially available influenza vaccine. In addition to an analysis of precursor frequency and light chain diversity, we studied sera and culture supernatants containing human anti-influenza antibodies with a panel of murine monoclonal antibodies specific for idiotopes identified on murine anti-PR8 and anti-X-31 antibodies. An idiotypic specificity present on the X-31-specific murine monoclonal PY206 has previously been shown to be shared by murine antibodies specific for PR8, X-31, and other influenza viruses. We observed little correlation among the following parameters: anti-viral titer, serum idiotope content, precursor frequency and immune status. More interestingly, there was a striking predominance of human influenza-specific antibodies that utilized lambda light chains. In addition, 12 of 26 human anti-influenza monoclonals strongly inhibited the binding of one of the murine anti-idiotopes to the labeled murine antibody, PY206. This is the same idiotope that is shared among murine antiinfluenza antibodies and all six individuals studied contained clones reactive with this anti-idiotope. Seven of these 12 idiotope-positive human antibodies gave partial cross-reactivity in a second anti-idiotypic system. These observations imply that a significant level of homology exists between the binding sites of human and murine influenza-specific antibodies and suggest that idiotypic manipulation of the human immune response to influenza virus may have important therapeutic implications.

Intravenous injection of New Zealand White rabbits with type III pneumococcal polysaccharide vaccine conjugated with the cardiac glycoside digoxin resulted in the production of both antidigoxin and anti-type III pneumococcal polysacharide antibodies. Among antisera of 12 rabbits examined during their peak antibody production periods, 1 to 20 mg (mean, 5.4 mg) of antidigoxin antibody could be recovered from 1 ml of serum. Antisera from five of these 12 rabbits contained antidigoxin antibodies of restricted heterogeneity as demonstrated by urea-polyacrylamide disc gel electrophoresis of fully reduced and alkylated antibodies. From the antisera of four of these five rabbits, electrophoretically homogeneous antibodies (1 to 5 mg/ml antiserum) could be isolated by affinity chromatography on ouabain-amine-Sepharose columns. The structural homogeneity of two of these antidigoxin antibodies was confirmed by amino acid sequence analysis of purified light chains through the first hypervariable region. These data suggest that the conjugation of small molecules to bacterial polysaccharide vaccines may provide a general method for synthesis of immunogens that can regularly elicit antihapten antibodies of restricted heterogeneity.

In order to examine the diversity of antibody light chains, we have developed an analytic isoelectric focusing procedure which permits the routine analysis of L chains from antibodies raised in individual mice. We have used this technique to demonstrate that the light chains of IgM and IgG anti-group A streptococcal antibodies raised in SWR mice are probably shared. Interestingly, numerous light chain spectrotypes are shared between individual mice whose 7S antibody focusing patterns differ.

The anti-phosphocholine (PC) memory response of BALB/c mice to PC-KLH contains two groups of antibodies distinguished by fine specificity and by expression of the T15 idiotype that dominates Group I but not Group II anti-PC antibodies. The contribution of V kappa genes to this diversity was investigated by the analysis of L chains from PC-binding hybridoma proteins (PCBHP) representative of Group I and Group II. N-terminal amino acid sequence analysis was performed on the L chains of three independently derived Group II PCBHP up to residue 23 (PCG1-1) or 21 (aPC-111-1 and aPC-12-3). These three sequences differed from each other by only one or two residues, but differed by approximately 50% from the L chains of the Group I-like PC-binding myeloma proteins (PCBMP); the Group II sequences are closely related to V kappa 1-3. Isoelectric focusing analysis was also performed on the L chain of PCG1-1, as well as on L chains from PCBHP typical of Group I antibodies, and from an atypical PCBHP differing from Groups I and II in fine specificity. A Group I PCBHP and the atypical PCBHP expressed L chains related to V kappa 8 and V kappa 24, respectively. The L chains of another Group I PCBHP and of the Group II protein, PCG1-1, appeared different from those found in the PCBMP and from each other. The results indicate a more diverse expression of L chains in the memory anti-PC response than is represented by the PCBMP; both V kappa 8- and V kappa 24-derived L chains (and, presumably, somatic variants), as well as products of additional V kappa genes (V kappa 1-3), appear to be present in the anti-PC memory pool.

The unreduced immunoglobulins (Ig) in the bullfrog, Rana catesbeiana, dissociate into two components when subjected to electrophoresis or molecular sieving in dissociating solvents. One of these components is monomeric light chain and the other is a disulfide-bonded complex of heavy chains. This unusual behavior has been observed with all classes of bullfrog Ig that have been isolated and characterized previously: a high m.w. Ig that resembles mammalian IgM and two antigenically distinct varieties of low m.w. Ig. Light chains, isolated from the high m.w. Ig by gel filtration in 8 M urea, 1 M acidic acid, were found to contain, on average, 5.7 residues of half-cystine. None of these residues were in the free sulfhydryl form nor were they blocked by half-cystine. Moreover, none was alkylated after mild reduction of the high m.w. Ig. These findings indicate that none of the light chain half-cystine residues participate in an interchain disulfide bridge, and that most of the light chains contain three intrachain bridges. This unusual pattern of disulfide bonding appears to be responsible for the noncovalent association of heavy and light chains in this species.

Murine myeloma immunoglobulin (IgA, K) J539, which shows enhanced tryptophanyl fluorescence on ligand binding, and S10, which shows reverse-sign changes in tryptophanyl fluorescence on ligand binding (RLIF, see below), have been reduced, alkylated, and dissociated into their light (L) and heavy (H) chains. Two hybrid recombinants, H10L539 and H539L10, have been prepared and the 7S material has been isolated by chromatography. The binding behavior of these recombinants was studied with a number of ligands. Both recombinants showed activity with beta(1 leads to 6) linked galactose ligands comparable to the native immunoglobulins. The ligand-induced fluorescence changes of the recombinants paralleled those of the heavy chain donor. For the recombinant H10L539, two different galactose-ligands caused fluorescence changes in opposite directions. It was quantitatively shown that binding of these ligands, nevertheless, took place in the same combining region. The idiotype of each recombinant resembled that of the heavy chain donor.