VH and VL gene elements that encode human antibodies to DNA

We have determined the cDNA sequence of variable regions of heavy and light chains of three antibodies with low affinity to DNA. The variable heavy chains were found to result from utilization of VH gene elements that have been identified previously in other low-affinity anti-DNAs. These VH gene elements, VH26 and VH1.9III, are expressed in association with different D gene-encoded CDR3s than in the other antibodies. The variable light chains were found to be encoded by VL gene elements that have not previously been identified in anti-DNAs. The recurrent identification of only 10 germ line VH genes in 22 low-affinity anti-DNAs indicates that there is a high probability that all of the VH gene elements that confer reactivity with DNA have been identified. Previous studies have suggested that high-affinity antibodies to DNA result from an antigen-driven process of affinity maturation. However, only 6 of 13 high-affinity antibodies to DNA are derived from this set of low-affinity VHs, indicating that DNA is unlikely to be the driving antigen.

Analysis of the individual contributions of immunoglobulin heavy and light chains to the binding of antigen using cell transfection and plasmon resonance analysis

We have cloned the Tg10 murine monoclonal antibody, which is specific for a human thyroglobulin (hTg) epitope targeted by autoantibodies in several thyroid pathologies. Transfection of COS-7 cells with plasmids expressing Tg10H and -kappa chains combined with surface plasmon resonance analysis (BIAcore) of culture supernatants showed that the entire cloned Tg10 antibody displays an affinity comparable to that of the parental antibody. This approach also permitted determination of the probable role of each chain to the recognition of the cognate epitope due to the ability of COS-7 cells to secrete independently each of the two constituting immunoglobulin chains. Tg10 heavy chain recognizes hTg in the absence of the light chain, but with a ten-fold lower affinity mainly due to an increase in kappaoff. In contrast, the light chain is unable to bind hTg on its own. This suggests that the latter is probably involved in stabilization rather than in initiating the formation of the antibody/antigen complex and that the specificity of Tg10 is mostly, if not exclusively, carried by the heavy chain. The potential applications of combined cell transfection and surface plasmon resonance to our understanding of antigen/antibody interactions are discussed.

A single-chain Fv reactive with the Goodpasture antigen

Goodpasture's disease is defined by the presence of autoantibodies to the glomerular basement membrane and characterized clinically by rapidly progressive glomerulonephritis and pulmonary hemorrhage. P1, a murine monoclonal antibody to the Goodpasture antigen (the noncollagenous domain of the alpha 3 chain of type IV collagen, alpha 3(IV)NC1), has been a valuable reagent in investigating the pathogenesis of this disorder. The purpose of this study was to generate and characterize a recombinant form of P1 as a single-chain Fv (scFv). First strand cDNA was made from RNA extracted from the P1 hybridoma cell line, and DNA encoding the antibody light and heavy chain variable domains was amplified by polymerase chain reaction, using universal oligonucleotides. The purified products were ligated sequentially into an expression plasmid separated by a sequence encoding a 15 amino acid flexible oligopeptide linker. The resulting scFv was expressed in E. coli. Functional scFv, designated HBR-3, was obtained by denaturing and refolding the expressed product. HBR-3 was shown by ELISA, immunoblotting, and immunohistologic techniques, to have the same specificity for alpha 3(IV)NC1 as P1 and autoantibodies from patients with Goodpasture's disease. HBR-3 and P1 were shown to have similar affinity for their mutual ligand. On sections of normal human kidney, the scFv bound only to glomerular basement membrane and distal tubular basement membrane. It did not bind to the glomerular basement membrane of patients with Alport's syndrome, in whom the Goodpasture antigen is often not expressed in an antigenic form. We have, therefore, generated a scFv which reproduces the specific binding properties of the parent monoclonal antibody, P1. The potential of HBR-3 as a diagnostic reagent in Alport's syndrome has been demonstrated. The development of this recombinant molecule should permit new approaches to the investigation of Goodpasture's disease.

The generation of antibody diversity in the turtle

The Ab response in reptiles has been studied at the protein level, and in turtles some aspects resemble those of cold-blooded vertebrates from other classes. The genetic bases for these features are not clear. The present study is the first on the IgH organization and complexity of a reptilian Ig gene system. The approach to cloning turtle (Pseudemys scripta) sequences is entirely PCR based, and its efficacy is demonstrated by obtaining extensive information on a heretofore unexplored Ig gene system. A number of genomic VH sequences, representing possibly four families, were isolated, as was a genomic C mu 4 clone. These sequences, used as probes, provided proof that in the turtle there is a single IgH locus with multiple VH genes and one C mu gene. In Northern hybridizations, the C mu 4 probe detected two transcripts; of the four VH groups, only one was expressed, and multiple bands indicated the presence of at least two non-mu transcripts. Using reverse transcription-PCR on spleen or liver RNA, an IgM heavy chain sequence was obtained, as were a number of VDJ rearrangements. Among 32 unique VDJ rearrangements from one animal, there were 22 sequence variants at framework 4, suggesting either a very large number of J segments or somatic modification in the variable region. The latter interpretation is supported by point mutations found in framework 3 and CDR3. The number of changes is considerably greater than the deduced Taq misincorporation rate (0.05%).

Four IgG anti-islet human monoclonal antibodies isolated from a type 1 diabetes patient recognize distinct epitopes of glutamic acid decarboxylase 65 and are somatically mutated

The selective destruction by an autoimmune process of the beta cells in the pancreas is the hallmark of the type 1 insulin-dependent diabetes mellitus. What triggers islet cell-specific autoreactive T and B cells, however, remains unclear. Identification of the targets of the anti-islet cell autoantibodies frequently found in insulin-dependent diabetes mellitus patients and analysis of their sequences should provide some insights into the nature of this disease. We have combined EBV transformation with CD40 activation of peripheral B cells from one patient with insulin-dependent diabetes mellitus to isolate four B cell clones that secrete islet cell-specific autoantibodies, These four human monoclonal autoantibodies are of the IgG1 isotype, and they each recognize a different epitope of the glutamic acid decarboxylase enzyme. Analysis of their variable gene sequences shows that, while clonally unrelated, three of the four human monoclonal autoantibodies use a member of the VH4 family, and two have rearranged the same delta light chain variable gene. The IgG1 isotype of the four autoantibodies as well as the presence of somatic mutations in both heavy and light chain genes provide concrete evidence for their derivation by a T cell-dependent immune response.

Humanization and molecular modeling of the anti-CD4 monoclonal antibody, OKT4A

OKT4A, a murine mAb that recognizes an epitope on the CD4 receptor, is a potent immunosuppressive agent in vitro and in a variety of nonhuman primate models of graft rejection and autoimmune disease. Initial human cardiac transplant trials suggest that OKT4A does not cause either cytokine release syndrome or CD4+ cell depletion, but does induce a human anti-mouse Ab (HAMA) response despite strong concurrent immunosuppression. To further investigate the potential of OKT4A as an immunomodulator, it was necessary to decrease its immunogenicity. Therefore, we developed a humanized version of this Ab (gOKT4A-4), which has the same binding affinity and in vitro immunosuppressive properties of OKT4A, but retains only three murine sequence-derived amino acid residues outside of the complementarity-determining regions (CDRs). Detailed computer modeling of both OKT4A and gOKT4A-4 provided a computational rationale for the changes necessary to regain activity after humanization. This has also provided a plausible representation of the Ag binding site. Preliminary clinical results with gOKT4A-4 suggest that we have eliminated the immunogenicity observed in the parent murine Ab.

The sheep Ig variable region repertoire consists of a single VH family

Nine germ-line Ig heavy chain variable (VH) segments (including three pseudogenes) were isolated from a genomic DNA library, and the other six were obtained by PCR, using 5'and 3' primers deduced from the first three. They appear to belong to a homogeneous VH gene family, with >80% sequence identity. This sheep VH gene family is related to the human VH4 family and to the murine VH1 subgroup (clan II). Southern blot analysis shows a maximum of 10 positive restriction fragments; therefore, the nine VH genes isolated probably constitute the major part of the repertoire. Thirty-one expressed mu variable regions (and one gamma 1 variable region) were obtained from adult spleen by either cDNA cloning or anchored reverse transcriptase-PCR; they are >80% similar to each other (in their leader to framework 3 regions) and to the germ-line sequences as well. The sheep VH repertoire thus seems to derive from a small (approximately 10 members) germ-line gene family, and its diversification must rely chiefly on junctional (D and/or N regions) diversity and somatic hypermutations.

A DNA repair abnormality specific for rearranged immunoglobulin variable genes in germinal center B cells

The somatic hypermutation mechanism produces high-rate mutagenesis specifically targeted to rearranged immunoglobulin (Ig) variable (V) gene segments during the germinal center (GC) stage of B lymphocyte differentiation. The mechanism of this process remains uncertain, partly due to the lack of a direct assay for hypermutation activity. In this study, a gene-specific DNA repair assay was used to compare the rate and quality of DNA repair in the mantle zone (MZ) and GC B cells at rearranged and unrearranged Ig V genes. GC B cells were distinguished from MZ B cells by a retarded repair rate specific for rearranged Ig V genes. In addition, a unique feature of GC cells after DNA repair was the appearance of predominant mutations in rearranged Ig VH5 gene PCR products. These predominant mutations also occurred in natural mutants of VH5 genes. However, repair-associated mutations reflected, at least in part, "template-jumping" during amplification of the residually damaged genomic template. Overall, these findings reflect a repair abnormality associated with the hypermutation process by the criteria of sequence- and B cell stage-specificity. We conclude that locus-specific retardation of DNA repair is a component of the hypermutation mechanism. RFLP or SSCP analysis provides a simple assay to monitor this repair abnormality as a surrogate biochemical marker for hypermutation during B cell differentiation.

The VDJ repertoire expressed in human preB cells reflects the selection of bona fide heavy chains

In early steps of B cell differentiation, mu chains are transiently expressed in association with a surrogate light chain (psi L) composed of the lambda-like and VpreB monomorphic polypeptides, thus forming a putative preB receptor. Using a monoclonal anti-VpreB antibody, preB cells were isolated from two adult human bone marrow samples and their VDJ repertoire analyzed at the transcription level. All VH families were identified and further analysis focused on VH3 sequence analysis of 37 distinct VDJ cDNA clones. The VH3 genes expressed in the two bone marrow samples were also encountered in fetal liver and adult peripheral blood lymphocytes with a roughly similar contribution of 3.30, 3.23, 3.9 and 3.53. The characteristic features of the preB repertoire as compared to the activation B repertoire include the quasi absence of somatic mutations, limited N diversity and a shorter third complementarity-determining region (CDR3). It also significantly differs from the fetal repertoire, which makes higher usage of DQ52 and has CDR3 of even shorter lengths. The almost constant presence of glycine residues in the CDR3 and predominance of JH4 with a low level of DQ52 DH usage, suggest that preB cell clones are submitted to an initial selective pressure which should be antigen independent. The bona fide heavy chains would be merely selected for their ability to interact with the surrogate light chains, thus shaping the repertoire that will be co-expressed with immunoglobulin light chains in IgM molecules.

cDNA sequence analysis and characterization of a cytokine-inducing monoclonal antibody derived from autoimmune MRL/MP-lpr/lpr mouse

We have previously reported that a monoclonal antibody 1D11 derived from an autoimmune MRL/Mp-lpr/lpr (MRL/lpr) mouse induced synthesis or increased production of IL-3, IL-6, and TNF-alpha in the IL-3-dependent bone marrow-derived cell line FDC-P2/185-4. In this report, we analyzed a sequence of cDNA encoding the V region of 1D11, and found that VH and VL segments of 1D11 belonged to the J558 and V kappa 21 family, respectively. The nucleotide sequence of 1D11 in the VH segment was highly homologous to that of AM9, a monoclonal RF derived from MRL/lpr mouse, and the only difference was the replacement of 3 nucleotides in the framework region 1 (FR1). However, the deduced amino acid sequence of 1D11 was identical to that of AM9. In contrast with the VH segment, the sequences of the VL regions of these two antibodies were quite different from each other; 1D11 showed a 3 base deletion in the FR2 and a 24 base insertion in the FR3 compared with AM9. At present, the mechanisms of such insertions or deletions in the FRs of autoantibodies are almost unknown. It is generally accepted that the differences in specificity and affinity of these autoantibodies depend on differences of CDR sequence. However, it is possible that not only CDRs but also FRs of autoantibodies play a critical role in pathogenicity and/or specificity in autoimmune diseases.

Characterization of the variable regions of a chimpanzee monoclonal antibody with potent neutralizing activity against HIV-1

The variable (V) regions of C108G, a potent neutralizing chimpanzee mAb against a glycan-dependent epitope in the V2 region of HIV-1 gp120, have been characterized for reactivity with human VH and VK family-specific antisera, and their nucleotide sequences have been determined and analysed. To our knowledge, this is the first study characterizing expressed chimpanzee VH and VK genes. Results show that C108G expresses members of the VH3 and VK1 families, the largest VH and VK families in humans, respectively. Nucleotide and amino acid sequence analyses reveal that C108G VH is most homologous to the human VH3 germline gene, hsigdp33 or V3-43, and the human JH4 minigene. The human germline VK1 gene that is most homologous to C108G VK, hsigk1012, was previously observed in unmutated form in a human autoantibody with anti-i red blood cell antigen specificity and in seven human Fabs and a mAb directed against epitopes overlapping the CD4-binding site of HIV-1 gp120. This germline gene was unmutated in three of the human Fabs and was somatically mutated in the other four Fabs and the mAb. In addition, the JK minigene was used in C108G VK, JK2, is apparently over-represented in anti-HIV-1 mAbs/Fabs; this minigene was used in 61% of the anti-gp120 human Fabs recently described and in three other anti-CD4-binding site human mAbs derived by EBV transformation. While the significance of these findings is unclear, they may suggest a bias in VK/JK gene usage and/or network regulation involving an hsigk1012/JK2 idiotope(s) in the antibody response to HIV-1. Both the C108G VH and VK genes showed evidence of somatic mutation and antigen selection that apparently occurred in vivo during chronic exposure to HIV-1 and its antigens. Surprisingly, this somatic mutation was most profound in the CDR3 region of C108G VK; this region shared only 48% nucleotide homology with hsigk1012 contrasted with a homology of 94% over the remainder of these two V gene sequences. Perhaps the most significant finding of this study is that the expressed VH and VK genes of chimpanzee mAb C108G are no more divergent from their most homologous human germline genes than are the expressed V genes of several recently characterized human anti-HIV-1 mAbs/Fabs from their apparent human germline genes. This suggests that chimpanzee mAbs are no more likely to elicit deleterious anti-immunoglobulin responses in humans than are human mAbs and emphasizes the potential for development of chimpanzee mAbs as immunotherapeutic agents.

Structural analysis of affinity maturation: the three-dimensional structures of complexes of an anti-nitrophenol antibody

Affinity maturation of the immune response to nitrophenol-containing antigens has been extensively investigated. Significant strides made during the past several years with the advent of PCR technology have provided a wealth of biochemical knowledge. Structural investigations of the phenomena have however been limited. We have determined the three-dimensional structure of the Fab fragment of 88C6/12, an anti-4-hydroxy-3-nitrophenyl acetic acid antibody complexed with the immunizing hapten and with a heteroclitic iodinated hapten. The crystallographic structure of the complexes reveals that the binding is stabilized by a number of hydrogen bonds and extensive van der Waals interactions between the hapten and the antibody. In addition, the Fab binding pocket contains a region of positive electrostatic potential well suited for interaction with the predominant resonance form of the nitrophenyl ring system. The observed heteroclicity towards the iodinated hapten is not a direct result of iodine-protein interactions, but results from the enhanced stability in the iodinated ring of the resonance form that binds the antibody. In addition this investigation provides a rationale for the strong preference for the substitution in the heavy chain from the germ-line gene encoded Trp 33 to Leu 33 in the mature anti-nitrophenol response.

Activity of a single-chain immunotoxin that selectively kills lymphoma and other B-lineage cells expressing the CD40 antigen

We have constructed anti-CD40 immunotoxins consisting of the single chain Fv (sFv) region of the anti-human CD40 mAb G28-5 fused to a truncated form of Pseudomonas exotoxin, PE40. CD40 is an integral membrane glycoprotein found on the surface of B-lineage cells, including lymphomas and leukemias, as well as certain carcinomas. Two forms of the immunotoxin were constructed, one with the light chain variable (VL) region of the sFv preceding the heavy chain variable region (VH) [G28-5 sFv(VL-VH)-PE40] and the second with the sFv in the opposite orientation [G28-5 sFv(VH-VL)-PE40]. Although both forms of G28-5 sFv-PE40 specifically bound to CD40 in ELISAs, the binding of G28-5 sFv(VL-VH)-PE40 was > 10-fold higher. A number of malignant B- and T-cell lines were screened for CD40 expression and susceptibility to G28-5 sFv(VL-VH)-PE40. All of the B-lineage cells tested were CD40 positive and sensitive to the anti-CD40 immunotoxin, with EC50s ranging from 2.5-70 ng/ml, whereas none of the T cell leukemias or lymphomas were antigen positive or were affected by the immunotoxins. Consistent with the antigen-binding results, the VL-VH immunotoxin was > 10-fold more cytotoxic than the VH-VL immunotoxin. The anti-CD40 single-chain immunotoxin fusion protein G28-5 sFv(VL-VH)-PE40 represents a potent and specific cytotoxic agent for the elimination of normal and transformed B-lineage cells expressing CD40.

Transferrin-antibody fusion proteins are effective in brain targeting

In the present study, the receptor binding potential of transferrin (Tf) was linked to an antibody binding specificity. Human Tf was fused to mouse-human chimeric IgG3 at three positions: at the end of heavy chain constant region 1 (CH1), after the hinge, and after CH3. The resulting Tf-antibody fusion proteins were able to bind antigen and the Tf receptor. The CH3-Tf fusion protein showed no complement-mediated cytolysis but possessed IgG receptor I (Fc gamma RI) binding activity. Most importantly, all of the fusion proteins demonstrated significant uptake into brain parenchyma, with 0.3% of the injected dose of the hinge-Tf fusion protein rapidly targeted to the brain. Recovery of iodinated CH3-Tf fusion protein from the brain parenchyma demonstrated that the fusion proteins can cross the blood-brain barrier intact. The binding specificity of these fusion proteins can be used for brain delivery of noncovalently bound ligands, such as drugs and peptides, or for targeting antigens present within the brain.

A subgroup of human VH3 germline genes that encode a high-avidity synovial rheumatoid factor

We have previously derived and identified a highly avid monoclonal IgM rheumatoid factor (mRF), C6, from unstimulated rheumatoid synovial cells (RSC). At the time, the closet VH germline gene, VH26, demonstrated only 88% homology with C6. To identify the germline counterpart of C6, genomic DNA from the same rheumatoid arthritis (RA) patient from whom C6 was derived was used in the polymerase chain reaction (PCR). Four of the six closely related germline genes that we sequenced had exonic regions that were identical with the VH region of C6 cDNA. These six germline sequences differed in their intronic regions, suggesting that they were distinct, but closely related genomic sequences. To further evaluate the extent of these related genes we identified nine additional germline genes having VH-encoding exons that were 86-97% identical to the C6 cDNA sequence. Furthermore, we examined the polymorphic nature of the C6 VH gene using single strand conformation polymorphism (SSCP), and identified two peaks, confirming the existence of highly homologous genes. The sequence and polymorphism data suggest that: (1) the VH region of the high avidity mRF C6 was derived from an unmutated germline gene; (2) C6 was encoded by a VH gene belonging to a set of homologous genes within the larger VH3 family; and (3) in addition to somatic rearrangements of B-cell genes and antigen-driven somatic mutation, gene duplication and conversion events of germline genes could be important in generating diversity and polyclonality among high-affinity pathogenic autoantibodies.

One disulfide bond in front of the second heavy chain constant region is necessary and sufficient for effector functions of human IgG3 without a genetic hinge

We have created four IgG3 mutants without a normal hinge region: (i) m0 without a genetic hinge; (ii) m0/C131S, where Cys-131 in m0 was mutated to Ser; (iii) m0/231C232 (formerly HM-1), where a Cys residue was inserted in m0 between Ala-231 and Pro-232; (iv) m0/C131S/231C232, which is a hybrid of m0/231C232 and m0/C131S. The wild-type IgG3 and all mutants bind 5-iodo-4-hydroxy-3-nitrophenacetyl groups. The wild type and mutants, m15 (with 15 aa in the hinge), m0/231C232, and m0/C131S/231C232, were all positive for complement-mediated lysis, antibody-dependent cellular cytotoxicity mediated by peripheral blood leukocytes, and phagocytosis by U937. m0/C131S/231C232 was only weakly positive and sometimes negative for respiratory burst activity mediated by peripheral blood neutrophils (polymorphonuclear leukocytes), whereas m15, m0/231C232, and wild-type IgG3 were strongly positive. The m0 and m0/C131S mutants were mainly negative for complement-mediated lysis, antibody-dependent cell-mediated cytotoxicity, and phagocytosis by U937 and polymorphonuclear leukocytes. The results indicate that a hinge spacer region is not necessary, but the correct alignment of the two second heavy chain constant regions in the IgG3 molecule by a minimum of one disulfide bond is necessary and sufficient for effector functions.

Fractionation of human serum proteins by capillary and recycling isotachophoresis

The fractionation of human serum proteins using capillary isotachophoresis (CITP) and recycling isotachophoresis (RITP) in presence of low molecular mass spacer compounds is reported. Anionic CITP was performed in an instrument equipped with a Teflon capillary of 0.5 mm ID as well as in an apparatus which features an open-tubular fused-silica capillary of 75 microns ID. RITP was performed in a recycling fast flow focusing apparatus in which fluid flows rapidly through a narrow, rectangular cell and the effluent from each outlet port is reinjected into the electrophoresis chamber through the corresponding input port. Typically, 1 mL serum was processed batchwise within about 2.5 h prior to collection of 30 fractions of about 4 mL each. The fractions were analyzed separately for conductivity, pH and UV absorbance and selected fractions were characterized by an immunoassay for transferrin, as well as by gel isoelectric focusing, two-dimensional gel electrophoresis, CITP and capillary zone electrophoresis. The search for suitable electrolyte systems and spacers was executed by CITP and by computer simulation. For simple configurations, separations predicted by simulation are shown to qualitatively agree with fractionation performed by CITP and RITP. Configurations producing three protein subgroups, the first containing mainly albumin, the second transferrin and the third the globulins, are discussed.

Protein-polysaccharide interactions. A monoclonal antibody specific for the capsular polysaccharide of Cryptococcus neoformans

Monoclonal antibodies that react with the capsular polysaccharide, termed glucuronoxylomannan (GXM), of Cryptococcus neoformans have potential roles in the diagnosis, monitoring of disease progress, and immunotherapy of cryptococcal GXM of all four serotypes. A molecular model of the Fab fragment of monoclonal antibody 439 was constructed from the amino acid sequence and a template antibody molecule, Fab 4-4-20. A tryptophan is present on the surface between light chain CDR3 and heavy chain CDR3 in the putative binding site. This tryptophan residue proved to be a reporter group, and a fluorescence study of Fab 439 was performed to analyze the interaction between cryptococcal GXM and Fab 439. Binding of the polysaccharide enhanced the intrinsic fluorescence and caused a blue shift in the emission maximum, indicating that the environment of a tryptophan changes from a polar to less polar environment. This is consistent with the loss of water from the binding site caused by the binding of antigen. This interpretation was confirmed by acrylamide quenching, which showed that 1 less tryptophan was exposed to solvent in the Fab-GXM complex than in free Fab. Fluorescence titration was used to determine binding and dissociation constants (KD). The apparent KD values for serotypes A-C were approximately the same; the KD for serotype D GXM was 5-11-fold greater. De-O-acetylation of serotype A GXM produced a 31-fold increase in the KD, indicating that the O-acetyl groups are important, but not essential, for binding. Carboxyl groups appear to be essential for strong binding because the KD for carboxyl-reduced GXM was so large that it could not be determined.

Crystallization of single-chain Fv proteins

Single-chain Fv (sFv) proteins consist of the variable heavy chain (VH) and variable light chain (VL) domains of an antibody, covalently joined by an engineered polypeptide linker. We report the crystallization of single-chain Fv's with specificities for fluorescein (4-4-20 sFv) and the TAG-72 pan-carcinoma glycoprotein antigen (CC49 sFv). Concentration of these proteins, preliminary to crystallization, results in a monomer-multimer equilibrium, causing aggregation which interferes with crystallization. Aggregation has been observed to depend primarily on an intact linker between VL and VH domains, although other factors are likely to modulate this phenomenon as well, including the specific identity of Fv and ligand, presence or absence of the ligand, linker length and possibly sequence. We have found two methods to overcome sFv aggregation, both of which yield X-ray diffraction quality crystals. The first, discovered serendipitously, is by introducing a proteolytic clip into the linker region (effectively yielding an Fv fragment). The second is the purification of the sFv dimer form, with linker regions intact, from an equilibrium mixture of aggregates. The sFv molecular association in a dimer is believed to be unusual in that each VL/VH interface may not be formed by the two linker-connected VL and VH domains, but rather by interaction of VL and VH domains from two distinct sFv monomers. Structure determination of the CC49 sFv dimer, with the 14-residue linker designated 212, is underway to test this model. Increasing linker length, to relieve steric strain on the monomer, and inclusion of the appropriate antigen, to slow transitions between monomeric and multimeric forms, may prove valuable strategies with sFv proteins less amenable to crystallization.

Human IgG preparations isolated by ion-exchange or protein G affinity chromatography differ in their glycosylation profiles

IgG from patients with rheumatoid arthritis (RA) is abnormally glycosylated in the Fc region, with sialic acid and galactose levels lower than normal. Protein G and DEAE purify populations which are differentially glycosylated. Significantly increased exposure of sialic acid was detected in normal IgG compared with that of RA IgG when ion exchange was used to prepare samples. However, when the same samples were prepared using protein G, no difference in the detection of sialic acid was seen between the two groups. When examining the heavy chain of IgG, more sialic acid, galactose and N-acetylglucosamine were detected in DEAE purified IgG compared with that prepared by protein G Detection of sialic acid and N-acetylglucosamine was also increased on light chains from IgG prepared by ion exchange chromatography. Since this occurs notably on rheumatoid light chains it would appear that this arrangement would contribute to the overall glycosylation changes in IgG. In the case of molecules lacking galactose the discrimination between the RA and normal IgG is significantly improved when ion exchange chromatography is used. Since differentiation between disease and normal groups relies on the purification technique used, we recommend that more than one method is employed before undertaking an analysis of glycosylation changes.

Purification and characterization of IgG1 and IgG2 from buffalo (Bubalus bubalis) serum and colostrum

Buffalo IgG1 and IgG2 were purified from serum and colostrum using salt precipitation, dialysis, gel filtration and ion-exchange chromatography. Their purity was monitored by immunodiffusion and immunoelectrophoresis using anti-heavy chain specific sera and SDS-PAGE. Selective binding of IgG2 to protein-A was used to remove IgG2 from IgG1 preparations. The IgG1 and IgG2 had a molecular mass (Mr) of 162.0 and 161.5 kD, respectively and were found to consist of heavy (H) and light (L) chains. The H and L chains had Mr of 58 and 24 kD, respectively. Reduction-alkylation followed by gel filtration was used for the isolation of H and L chains. While intact H chains were obtained, the L chains appeared to be cleaved into 14 kD molecules and smaller fragments. The mean hexoses content of the serum IgG1 and IgG2 was 1.81 +/- 0.02% and 0.70 +/- 0.02%, respectively. The corresponding values for colostral IgG1 and IgG2 were 1.76 +/- 0.01% and 0.78 +/- 0.08%. Both the IgG subclasses activated homologous complement. These results suggest that buffalo and cattle IgG subclasses have many common characteristics and minor differences.

Recombinant immunotoxins containing the VH or VL domain of monoclonal antibody B3 fused to Pseudomonas exotoxin

We prepared recombinant immunotoxins in Escherichia coli in which the VH or VL domains of mAb B3 were fused to a truncated form of Pseudomonas exotoxin (PE) (PE38KDEL). mAb B3 binds to a carbohydrate Ag found on the surfaces of many types of cancers and only a few normal tissues. PE38KDEL is a 38-kDa form of PE (66 kDa) that is missing the cell-binding domain of PE and has the carboxyl end changed from REDLK to KDEL. We show that immunotoxins in which the H chain or the L chain V region is fused to PE38KDEL bind to and kill carcinoma cells containing the B3 Ag. B3 Ag-negative cells were not affected. The cytotoxicity of these molecules is between 20- and 100-fold less than B3(Fv)-immunotoxins, containing both the H and L chain V regions. The VL-containing toxin was more active than the VH-containing toxin, indicating that the L chain of mAb B3 probably contributes more to Ag-binding than the H chain. Refolding experiments show that B3(VL)-PE38KDEL aggregates less than the VH-derivative or than a single chain immunotoxin B3(Fv)-PE38KDEL, which contains both domains in a single chain form. Furthermore, in addition to monomers, active homodimers of B3(VH)- and B3(VL)-PE38KDEL were obtained from renaturation experiments. The VL-toxin dimers, which might have their binding regions arranged in a manner similar to Bence Jones proteins (L chain homodimers), were found to have almost the same cytotoxicity as the monomers, whereas the VH-toxin dimers had decreased cytotoxic activity.

Pattern variations of polyclonal and monoclonal immunoglobulins of different isotypes analyzed by high-resolution two-dimensional electrophoresis

High-resolution two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) was used to analyze serum samples and purified immunoglobulins (Ig) obtained from "normal" individuals and from patients diagnosed with monoclonal gammopathies (MG) (n = 47; 5 IgA, 15 IgM, 15 IgG, 4 biclonal IgG, 1 IgD, 7 Bence Jones proteins). Polyclonal and monoclonal heavy (H) chains were located at different restricted gel positions according to their isotype. Monoclonal H chains appeared as sets of spots characterized by charge (pI) and size (M(r)) microheterogeneity. Most of the monoclonal gamma chains were not seen on the gels (12/15). Supplementary polypeptides of 45-48 kDa were detected in serum samples containing monoclonal IgM, but were not seen in MG of other isotypes. However, these polypeptides were not specifically associated with monoclonal IgM because they were also found on protein maps of purified polyclonal IgM. Polyclonal light (L) chains appeared as cloudy bands containing several zones of higher density, whereas monoclonal L chains were usually resolved as single sharp spots. In 6 samples, monoclonal L chains were not seen, and in 9 samples, they appeared as two or more spots, characterized by different pI and/or M(r). In one sample obtained from a patient with a biclonal gammopathy, the L chains were resolved as 4 different spots. Our results confirm that 2-D PAGE is an excellent tool to study Ig. Analysis of the L chain region of the gels was particularly informative. Several monoclonal L chains exhibited heterogeneous two-dimensional spot patterns, suggesting that "subtile" clonal mutations of B-cell lineage and/or posttranslational modifications were involved in their production.