The Primary Antibody Repertoire of κ-Deficient Mice is Characterized by Non-Stochastic Vλ1 + VH Gene Family Pairings and a Higher Degree of Self-Reactivity

We have investigated the primary antibody repertoire of genetically manipulated 129/Sv κ-deficient (JC_κD) mice, in order to understand the contributions of the λ-light chain, in the absence of an otherwise predominant κ-light chain, to the development of humoral immunity. The expression of V_λ1 gene (λ₁ and λ₃ subtypes) and the V_λ1 + V_H (J558, 36–60, V_H11 and S107) gene family associations were studied in 7.43 × 10³ mitogen-activated splenic B-lymphocyte clones of JC_κD origin. Furthermore, the functional significance of the exclusive expression of the λ-light chain, in the peripheral B-cell repertoire of JC_κD mice, was analysed by determining natural autoantibody specificities in the circulating serum immunoglobulin and the frequency of autoreactive B-lymphocyte clones in the peripheral B-lymphocyte repertoire. These experiments revealed that: first, of the three available V_λ genes at the λ locus, the V_λ1 gene is the one that is expressed most frequently (59.9%); second, non-random V_λ1 + V_H (J558, 36–60) gene family pairings occur in κ-deficient mice; and third, a higher degree of self-reactivity is generated as a result of exclusive use of the λ-light chain, as evidenced by higher levels of serum natural autoantibodies as well as a high frequency of autoreactive B-lymphocyte clones in κ-deficient (129/Sv JC_κD) mice. These observations suggest that the high murine κ/λ ratio in mice may, apart from high sequence diversity at the κ-locus, be a result of endogenous selection against the λ-light chain to restrict self-reactivity within the homeostatic threshold.     

Molecular analysis of rheumatoid factor (RF)-negative B cell hybridomas from rheumatoid synovial tissue: evidence for an antigen-induced stimulation with selection of high mutated IgVH and low mutated IgVL/λ genes

The mutational pattern of IgV_H and IgV_L genes from synovial tissue B cell hybridomas (n = 8) of patients (n = 4) with rheumatoid arthritis (RA) was analysed, which had been produced by the electrofusion technique without prior in vitro stimulation. The molecular data were correlated with immunohistopathological data and parameters of local disease activity. The IgV_H genes of the B cell hybridomas belonged to the V_H3 family (DP42; DP47, n = 2; DP53), the V_H1 family (DP75), the V_H4 family (DP71) and the V_H5 family (DP73); 7/7 IgV_H genes showed somatic mutations, the R/S ratio (CDR) was > 3 in 4/7 IgV_H genes and the mean R/S ratio of all IgV_H genes was 9.3 (CDR) and 1.0 (FR), suggesting an antigen-dependent selection. The IgV_L/λ genes belonged to the Vλ1 family (DPL2, DPL5, DPL8nf), the Vλ2 family (DPL11, n = 2) and to the Vλ6 family (IGLV6S1); 6/6 IgV_L genes showed somatic mutations, the R/S ratio (CDR) was > 3 in 3/6 IgV_L genes and the mean R/S ratio of all IgV_L was 3.0 (CDR) and 2.3 (FR), suggesting an antigen-dependent selection. The synovial tissue exhibited germinal centres in the follicles (3/4), with the unique distribution of Ki-M4⁺ follicular dendritic cells and Ki-67⁺ proliferating cells and a dominance of IgA⁺ plasma cells (3/3). All patients were positive for RF in serum and exhibited severe local symptoms (swelling 4/4; warmth 4/4; effusion 2/4), whereas the hybridomas were negative for RF. Since B cell hybridomas showed hypermutation and affinity selection for IgV_H and IgV_L/λ genes and the patients exhibited severe local symptoms with germinal centres in synovial tissue, this study indicates that an antigen-driven process is behind the B cell expansion in the synovial tissue of clinically affected joints. These mutated B hybridomas were negative for RF, thus suggesting that antigens different from RF are also involved in the local B cell expansion and in the chronic synovitis of RA.     

Variable region sequences of autoantibodies from mice with experimental systemic lupus erythematosus

We have sequenced nine monoclonal antibodies (mAb) derived from C3H.SW mice in which experimental systemic lupus erythematosus (SLE) was induced. The hybridomas were selected for binding to DNA or to HeLa nuclear extract (NE). Three mAb were found to bind DNA, and are shown to exhibit sequence characteristics of pathogenic anti-DNA antibodies. One, mAb 2C4C2, is shown to use a heavy chain V region gene (V_H) identical to the V_H of anti-DNA mAb isolated from other lupus-prone mice, namely (NZB × NZW)F₁. The light chain V region gene (V_L) of mAb 2C4C2 is 98 % homologous to the V_L of another anti-DNA mAb, also isolated from (NZB × NZW)F₁ mice. The other two anti-DNA mAb, 5G12-4 and 5G12-6, share 93 % of their V_H sequences with that of mAb 2C4C2. Six mAb bound proteins of HeLa NE. Four of these six antibodies were found to use the VH124 V_H and V-L7 V_L. The nine mAb use a total of five V_H and four V_L germ-line genes, demonstrating that the autoantibodies induced in mice with experimental SLE do not originate from one B cell clone. Three of these nine V_H and V_L were identical in sequence to germ-line genes, while at least three others had somatic mutations. The latter suggests that the above autoantibodies arise in mice by both usage of existing (pre-immune) B cells, and through an antigen-driven process. Furthermore, it appears that autoantibodies found in mice with experimental SLE use genetic elements similar to those used by mAb that were isolated from mouse strains which develop lupus spontaneously.     

Light chain variable (VL) sequences of rheumatoid factors (RF) in patients with primary Sjögren's syndrome (pSS): moderate contribution of somatic hypermutation

We have characterized and sequenced the variable (V) region genes of the light (L) chains of 10 immunoglobulin (IgM) rheumatoid factor (RF) monoclonal antibodies (MoAb) derived by the hybridoma/Epstein-Barr virus (EBV) technique from the peripheral blood of patients with primary Sjögren's syndrome (pSS). Six out of 10 RFs used lambda (λ) L chains, while four RFs used kappa (κ) L chains. Five out of the six λ RFs were encoded by V_λ3 gene segments, the sixth one was encoded by a V_λ1 gene segment. This preferential utilization of the V_λ3 family genes suggests selective expansion of the B cell in pSS. Three of the κ RFs used V_κ3 gene segments, while the fourth used a V_κ2 gene segment. Half of the RFs were found as unmutated copies of their closest germline (GL) gene. Interestingly these RFs were previously shown to use heavy (H) chains in GL gene configuration. Three RFs have very few mutations (2–3) and only two RFs have substantial numbers of mutations (6 and 11). This also correlated with the number of mutations in the respective H chains. In contrast to RFs in normal and RA these results further suggest the somatic mutation to be of moderate importance in the generation of RF from the peripheral blood of pSS patients.     

Preferential VH/V lambda pairings occur in the available B cell repertoire of adult BALB/c mice.

To gain insights into the composition of the B cell repertoire, we have investigated VH gene family expression associated with individual light chains. For this purpose, we have examined the use of 12 VH gene families in a large collection of hybridomas expressing one of the four lambda light chains [lambda 1 (V1J1), lambda 2 (V2J2 and V x J2) and lambda 3 (V1J3)]. Our results show that the distribution of the VH families is very different from one lambda subtype to another. This suggests that a few substitutions between VL regions are sufficient to generate very different associated repertoires by strong selection mechanisms. Moreover, we assume that the global VH expression pattern is not random but rather composed of many preferential VH/VL associations.     

The analysis of organization and diversity mechanism in goat immunoglobulin light chain gene loci

The genomic organization of goat immunoglobulin light chains (Igλ and Igκ) loci were annotated based on the goat genome database. The goat Igλ chain located on chromosome 17 contains at least 35 V_λ gene fragments (seven potential functional genes, one ORF and 27 pseudogenes), two J_λ-C_λ clusters arranged in a V_λ(35)-J_λ2-C_λ1-J_λ1-C_λ2 pattern, with another C_λ3 on scaffold. The Igκ locus included 11 Vκ (five potential functional genes, two ORFs and four pseudogene fragments), three J_κ genes and a single C_κ gene ordered in V_κ(35)-J_κ(3)-C_κ pattern on chromosome 11. By analyzing the clonies of Igλ and Igκ, we further found V_λ2 (26.23 %) & V_λ3 (73.11 %), V_κ2 (52.07 %) & V_κ4 (46.15 %) were predominately used in the expression of λ and κ chains respectively. λ chain showed more abundance in connective diversity than κ chain. Besides, somatic hypermutation with higher frequency in both immunoglobulin light chains was the major mechanism for the goat repertoire diversity. These results demonstrated goat immunoglobulin light chain variable region genome loci and repertoire diversity.     

Rabbit DQ52 and DH gene expression in early B-cell development

Rabbits predominantly rearrange the most 3′ V_H gene (VH1); thus combinatorial diversity is very limited. In man and mouse, the most 3′ D_H gene, DQ52, is preferentially rearranged early in B-cell development. To test whether this preference for rearranging a D_H gene segment based on 3′ end proximity exists in rabbit, we cloned and sequenced the rabbit DQ52 gene. The 11 base pair coding region sequence is identical to a published mouse DQ52, and 81.8% similar to the human sequence. It is localized 805 bp upstream of the J_H1 gene. However, the 3′ recombination signal sequence has an atypical nonamer. We prepared mRNA from 15- to 28-day fetal rabbits and amplified expressed VDJ sequences of μ mRNA by RT-PCR. The PCR products with VDJ rearrangements were cloned and sequenced. As expected, 44 of 45 VDJ sequences reflected use of the 3′ V_H1a2 gene, but the DQ52 gene was utilized very infrequently, if at all. We found only one VDJ sequence from 28-day fetal liver B-cells with 8 bp that matched the germline DQ52 sequence. Instead of expressing DQ52, another D_H gene, Df was frequently expressed. We cloned the genomic Df gene and localized it about 32 kb upstream of the J_H region. Thus, in contrast to man and mouse, rabbits preferentially express a D_H gene located in the middle of the D_H region early in B cell ontogeny. This may correlate with more frequent initial rearrangement of V_H to D_H in rabbit B cells.     

Pairing of VH gene families with the λ light chain: Evidence for a non-stochastic association

The frequencies at which four V_H gene families pair with the λ1 light (L) chain were determined by sequential hybridization of V_H- and λ1-specific DNA probes to mitogen-induced colonies of B cells. Analysis of pair frequencies indicates that the repertoire of λ L chain antibodies is generated by the stochastic pairing of smaller 3′-to-mid-locusV_H gene families (X-24, S107, Q52). However, the large 5′ V_H J558 family appeared to associate with the λ1L chain non-stochastically; the frequency of VhJ558/λ1⁺ colonies among all λ1⁺ colonies was significantly lower than the frequency of J558 expression among all (Cμ⁺) B cell colonies. This difference suggests that selection, either intrinsic at the level of rearrangement or heavy and L chain pairing, or extrinsic following surface immunoglobulin expression, may operate to shape the λ antibody repertoire prior to the introduction of exogenous antigen.     

VH gene-family representation in peripheral activated B cells from systemic lupus erythematosus (SLE) patients

A semiquantitative polymerase chain reaction (PCR) assay described in this study has been used to analyse the V_H1, V_H3 and V_H4 repertoire expressed by total IgM⁺ and IgG⁺ B cells from normal individuals and lupus patients. This approach consists of a combination of B cell selection, utilization of the anchored PCR technique to avoid technical bias in the amplification of different V_H gene family cDNA templates, and screening of the amplified IgM or IgG cDNA rearrangements by family-specific oligonucleotide probes. In four lupus patients, V_H family representation in IgM⁺ and IgG⁺ in vivo activated B cells, selected by anti-CD71 antibody, and in total CD19⁺ B cells were compared. In all patients, V_H4 gene family segments were preferentially underrepresented in IgM⁺ activated B cells. In IgG⁺ B cells the results suggest that V_H4 expression is variable, depending on the phase of the disease. Polyclonal B cell activation, which is usually considered as being the first event in autoantibody production in SLE, cannot explain our results. The data evoke the possible involvement of a V_H4-specific B cell superantigen in the onset or development of SLE. This hypothesis is also supported by the sequence conservation of the fourth β loop—a putative superantigen binding site—of functional V_H4 gene segments which are preferentially used by anti-dsDNA lupus antibodies of established clones and hybridomas.     

Preparation and characterization of antibodies to the lambda chain variable region (Vlambda) of mouse immunoglobulins.

The variable portion (V_L) of the mouse myeloma protein MOPC-315 (α, λ₂) was obtained from its Fv fragment and was used to immunize rabbits. Anti-V_L antibodies were found to be specific to the V_L region of protein 315 and to precipitate V_L, Fv or light chain but not Fab' or intact M315. Anti-V_L 315 precipitates also the light chain of MOPC-104E (μ, λ₁) but not the intact protein. Intact immunoglobulins (Ig) containing λ chains, although they do not precipitate with anti-V_L, inhibit the binding of anti-V_L to V_L in the radioimmunoassay. Radioimmunoassay inhibition studies demonstrated that V_λ1 and V_λ2 cross-react extensively and that anti-V_L 315 can be used as a general anti-V_λ reagent to detect molecules containing λ chains in mouse serum. Analysis of sera from several mouse strains indicated that V_λ-containing molecules are present in approximately 3% of the Ig population, whereas the SJL strain has no detectable V_λ-bearing molecules. In some of the antisera, anti-V_L 315 slightly cross-reacted with V_K-bearing molecules. Anti-V_L 315 antibodies are not anti-idiotypes since they react with myeloma proteins M315 (anti-2,4-dinitrophenyl MOPC-104E (anti-dextran) and HOPC-1 (specificity unknown) and with most serum Ig containing λ chains. In view of the similarity of mouse λ chain sequences, it is suggested that most anti-V_L 315 is “anti-framework” antibody.     

Development of the early B cell population in Xenopus

Like mammals, the amphibian Xenopus uses combinatorial joining of the immunoglobulin V, D and J elements and multiple rearrangements to generate its B cell repertoire. Xenopus larvae hatch 2 days after fertilization and individuals are under pressure to develop an immune repertoire when the number of available cells is small (approximately 5 and 200 IgM-positive cells on days 5 and 11 after fertilization, respectively). In the liver, in a first phase of differentiation spanning days 5 – 12 after fertilization before immunological competence, the heavy (H) chain locus starts rearranging followed by the light (L) chain locus 3 days later. By immunohistology the first B cells expressing H and L chain are detectable on day 10. Despite the small number of cells available and the lack of external antigen selection at these early stages, the repertoire is heterogeneous. The V_H families are used stepwise, although their genes are interspersed in the genome. The earliest family used (V_H1) is homologous to the V_H3 family of human and to the V_H7183 of the mouse which are also overrepresented in early mammalian development. In the second phase, from day 12 – 13 onwards, the spleen differ entiates and the animal becomes immunologically competent. The V, D and J usage is similar to that of adults although VDJ junctions lack N nucleotides until metamorphosis. A preferential reading frame for D and one specific DJ junction are overrepresented during this second phase. The visible bias toward homology-based junction results in fact from selection after rearrangement.     

Biased usage of two restricted VH gene segments in Vh replacement

AT8-1-12-5, an intracytoplasmic γ2b-producing (μ-γ2b⁺) pre-B cell line transformed with Abelson murine leukemia virus continuously generated intracytoplasmic μγproducing (μ⁺μ2b⁺ cells during propagation in culture. Southern blotting, DNA cloning and sequence analysis showed that these μ. ⁺μ2b⁺ cells were generated from the μ- γ2b⁺ pre-B cells by V_H replacement events. The gene segments involved in these replacement events were restricted to only two V_H gene segments, V^h7.1 (V^h7183 family) and V^h6.2 (V^hQ52 family). Cell staining using the monoclonal anti-V^H141 antibody 3-5-6f that specifically recognized the V_H6.2 but not V_H7.1 gene products indicated that half of the V_H replacement events occurring in AT8-1-12-5 used the V_H6.2 gene segment. Deletion mapping indicated that the incoming V_H gene segments, V_H7.1 and V_H6.2 were proximal to the resident V_H gene segment, V_H12.5 (V_H7183 family). These results provided direct evidence of the biased usage of VH gene segments in VH replacement and have several implications for the mechanisms of V_H replacement.     

Immunoglobulin VH region genes of the mouse are organized in overlapping clusters

Restriction fragment length polymorphism has been compared between the Igh loci of C57BL/6 and MOLF/EI (Mus musculus molossinus) mice utilizing probes which detect the C_γ2b gene and genes from nine V_H-gene families. Distinct restriction site patterns were found for the CH genes and for V_H families PC7183, Q52 and X24. V_H families V31 and J558 showed identical patterns. Mixed patterns of identical and distinct bands were detected in V_H families S107, J606, V3660 and VGAM3.8. This indicates that a recombination took place involving the Igh loci of a M. m. molossinus and a progenitor of the C57BL/6 strain. The breakpoint of recombination maps to the chromosomal region carrying V_H families S107, J606, VGAM3.8 and V3660. V^H families PC7183, Q52 and X24 map 3′ to the recombination breakpoint and proximal to the D_H-J_H-C_H region, whereas V_H families V31 and J558 accordingly map 5′ to the recombination breakpoint and distal to D_H-J_H-C_H. This order of V_H families was confirmed by deletion mapping utilizing hybridomas which are haploid either for the Igh^b or for the Igh^a locus. The mapping data indicate that the V^H families of the mouse are organized in overlapping clusters. This notion is confirmed by demonstration of the physical linkage of V_H genes from families V31 and J558 in the Igh^a locus.     

The anti-La response of a single MRL/Mp-lpr/lpr mouse: Specificity for DNA and VH gene usage

Autoantibodies to ribonucleoproteins (RNP) occur prominently in human systemic lupus erythematosus and murine lupus models. In previous studies we demonstrated a relationship in MRL/Mp-lpr/lpr (MRL/lpr) mice between antibodies to Sm, an RNP autoantigen, and antibodies to DNA. Thus, many anti-Sm monoclonals bound DNA and expressed the same V region genes as anti-DNA. In addition, many had multiple V_HCDR3 Arg residues suggestive of selection by DNA, and some had somatic mutations suggesting selection for mutant B cells by DNA. To determine whether autoantibodies to other RNP antigens are also associated with the anti-DNA response, we have analyzed the response to the La RNP. Six anti-La B cell hybridomas were generated from a single MRL/lpr mouse. Southern blot analysis of Ig V gene rearrangements and V gene sequences indicated two clonally related pairs, suggesting an oligoclonal response. Antibodies from all six hybridomas bound single-stranded DNA, while antibodies from five hybridomas bound double-stranded DNA. Two hybridomas expressed a V_H7183 gene which is used by members of two previously reported anti-DNA clones and two anti-Sm/DNA clones of MRL/lpr origin. These data demonstrate an association between the anti-La and anti-DNA responses in MRL/lpr mice, suggesting that cross-reactive anti-RNP and anti-DNA responses are a general feature of autoimmunity in this lupus model.     

Analysis of individual immunoglobulin λ light chain genes amplified from single cells is inconsistent with variable region gene conversion in germinal-center B cell somatic mutation

Responding B cells in specific immune responses diversify their immunoglobulin genes and are selected on their variant antigen receptors in the microenviroment of the germinal center. The patterns of mutations previously reported for immunglobulin (Ig) genes have supported mechanistic hypotheses of either error-prone DNA synthesis or templated variable region gene conversion as the underlying mechanism in the generation of these mutations. To assess the role of gene conversion in germinal-center somatic mutation, we chose to examine nucleotide changes in mouse λ light chain genes which arose in response to a specific antigen. Laboratory mice possess three Vλ subexons, two of which differ from one another by only seven nucleotides, making these two subexons ideal for gene conversion. In the current study, we used six-parameter flow cytometry to isolate single λ light chain-expressing germinal-center B cells from two different time points in a primary immune response. We then individually amplified and sequenced individual V_λ1 genes from these single cells for mutational analysis. None of the 32 V_λ1 genes, containing a total of 40 mutations, showed evidence of gene conversion from either of the other Vλ subexons. Features such as the replacement to silent ratio of the mutations documented at the earlier time point indicate an absence of antigen-driven selection. These data indicate that V region gene conversion does not contribute to germinal-center somatic mutation and that gene conversion is not responsible for targeting mutation specifically to rearranged Ig genes. The biological implications are discussed.     

B lymphocyte recognition of cytochrome c: higher frequency of cells specific for self versus foreign antigen early in the immune response and V gene usage in the response to self antigen

To study immunoglobulin gene usage in the antibody response of mice to the self antigen (Ag) mouse cytochrome c (cyt), B cell hybridomas were prepared from splenic B cells of immunized BALB/c mice prior to the onset of somatic mutation, i.e. 3 days after injecting ovalbumin (OVA)-primed mice with mouse cyt coupled to OVA. Monoclonal antibodies (mAb) from all of the seven primary hybridomas we obtained were sensitive to a single amino acid substitution from aspartic acid to glutamic acid at position 62 in mouse cyt. This is the specificity of the vast majority of B cells responding to mouse cyt as determined from assays of B cells activated in splenic fragment cultures. Six of the mAb derive from the 19.1.2 J558 V_H gene which is also used in the response to α (1 → 6) dextran and three of these mAb derive from the R9 V_x gene, a member of the V_xOx-1 family. The other mAb derive from distinct, although similar, V_x genes. Attempts to obtain hybridomas secreting primary (unmutated) mAb specific for cyt foreign to mice have been hampered by the much lower frequency of B cells responding early to foreign cyt in comparison to the self Ag. This suggests that, contrary to expectation of tolerance mechanisms, in naive BALB/c mice B lymphocytes specific for a single epitope on self cyt are present in higher frequency than B lymphocytes specific for similar epitopes on foreign cyt. Possible explanations for this result include biased expression in the B cell repertoire of the particular combination of V genes encoding mouse cyt-specific mAb or to positive selection of developing B lymphocytes by endogenous Ag.     

VH usage and somatic hypermutation in peripheral blood B cells of patients with rheumatoid arthritis (RA)

The human antibody repertoire has been demonstrated to have a marked V-gene-dependent bias that is conserved between individuals. In RA patients, certain heavy chain V genes (V_H) have been found to be preferentially used for encoding autoantibodies. To determine if such preferential use of V_H genes in autoantibodies is associated with a general distortion of the V gene repertoire in RA patients, the V_H composition of peripheral blood B cells was analysed among four RA patients and four age- and sex-matched healthy controls. Usage of individual V_H genes (eight V_H3 and three V_H4 genes tested by hybridization with a set of gene-specific oligonucleotide probes) was highly biased among RA patients, but no evidence of a distortion in the bias was observed compared with healthy controls. However, the occurrence of somatic mutations in these V_H genes (estimated by differential hybridization with motif-specific oligonucleotide probes targeted to CDR and FR of the tested genes, and by DNA sequence analysis) was strikingly different between patients and healthy subjects. The number of V_H3 rearrangements that had accumulated somatic mutations and the number of mutations per rearrangement were significantly elevated in three of the four RA patients. A slight but not significant elevation in mutations among rearranged V_H4 genes was also observed in these patients. These data suggest that although usage of individual V_H genes among peripheral blood B cells is not affected by the disease, the autoimmune process may involve a significant fraction of the B cell compartment.     

Preferential use of the VH5-51 gene segment by the human immune response to code for antibodies against the V3 domain of HIV-1

Human anti-V3 monoclonal antibodies (mAbs) generated from HIV-1 infected individuals display diversity in the range of their cross-neutralization that may be related to their immunogenetic background. The study of the immunoglobulin (Ig) variable region gene usage of heavy chains have shown a preferential usage of the VH5-51 gene segment which was detected in 35% of 51 human anti-V3 mAbs. In contrast, human mAbs against other envelope regions of HIV-1 (anti-Env), including the CD4-binding domain, the CD4-induced epitope, and gp41 preferentially used the VH1-69 gene segment, and none of them used the VH5-51 gene. Furthermore, the usage of the VH4 family by anti-V3 mAbs was restricted to only one gene segment, VH4-59, while the VH3 gene family was used at a significantly lower frequency by all of the analyzed anti-HIV-1 mAbs. Multivariate analysis showed that usage of VH gene segments was significantly different between anti-V3 and anti-Env mAbs, and compared to antibodies from healthy subjects. In addition, the anti-V3 mAbs preferentially used the JH3 and D2-15 gene segments. The preferential usage of selected Ig gene segments and the characteristic pattern of Ig gene usage by anti-V3 mAbs can be related to the conserved structure of the V3 region.     

Synovial IgG rheumatoid factors show evidence of an antigen-driven immune response and a shift in the V gene repertoire compared to IgM rheumatoid factors

We have established IgG rheumatoid factor (RF)-secreting hybridoma cell lines from the synovial tissues of three patients from whom we have previously characterized several IgM RF. The IgG RF bind human and rabbit IgG and form intracellular complement-fixing complexes indicative of a self association process in vivo. Nucleotide sequence analysis revealed that two IgG RF used V_HIII gene segments, while one used a V_HI gene segment. The V_L gene usage consisted of a Vχ1, a Vλ2 and a Vχ4/Vχ6 hybrid, confirming our previous findings that many different V_L genes can contribute to RF specificity. Although the IgG RF used V_H genes from the same families that dominated the IgM RF response, two of the actual gene segments employed were not found among the IgM RF. In contrast to IgM RF, which in general were not very mutated, the IgG RF showed somatic mutations characteristic of an antigen-driven immune response.