Parallel evolution of antibody variable regions by somatic processes: consecutive shared somatic alterations in VH genes expressed by independently generated hybridomas apparently acquired by point mutation and selection rather than by gene conversion

The bovine genomic DNA sequence data reveal three IGHV subgroups, only one of which is functionally expressed

A comprehensive analysis of cattle shotgun sequencing data reveals 36 immunoglobulin heavy chain variable genes. The previously described bovine subgroup IGHV1 contains 10 functional genes with a conserved promoter including the consensus octamer and several other transcription factor binding sites, intact exons and matching cDNA sequences. Subgroups IGHV2 and IGHV3 consist entirely of pseudogenes. Thus, the bovine germline IGHV repertoire is very limited. The IGHV genes are distributed in mammalian clans I and II, while no clan III genes were detected. Clan-specific PCR of genomic DNA from cattle, sheep, Eurasian elk, white-tailed deer, pig and dolphin indicates highly dynamic evolution of IGHV gene usage within Cetartiodactyla. The bovine germline IGHV repertoire was probably generated by recent duplications of an IGHV1-IGHV2 homology unit. Immunoglobulin heavy chain genes are largely incorrectly assembled in the current cattle genome versions Btau_4.2 and UMD_3.1. FISH experiments confirm an IGHV locus close to terminus of BTA21.

Gene Conversion-Like Events in the Diversification of Human Rearranged IGHV3-23*01 Gene Sequences

Gene conversion (GCV), a mechanism mediated by activation-induced cytidine deaminase (AID) is well established as a mechanism of immunoglobulin diversification in a few species. However, definitive evidence of GCV-like events in human immunoglobulin genes is scarce. The lack of evidence of GCV in human rearranged immunoglobulin gene sequences is puzzling given the presence of highly similar germline donors and the presence of all the enzymatic machinery required for GCV. In this study, we undertook a computational analysis of rearranged IGHV3-23(*)01 gene sequences from common variable immunodeficiency (CVID) patients, AID-deficient patients, and healthy individuals to survey "GCV-like" activities. We analyzed rearranged IGHV3-23(*)01 gene sequences obtained from total PBMC RNA and single-cell polymerase chain reaction of individual B cell lysates. Our search identified strong evidence of GCV-like activity. We observed that GCV-like tracts are flanked by AID hotspot motifs. Structural modeling of IGHV3-23(*)01 gene sequence revealed that hypermutable bases flanking GCV-like tracts are in the single stranded DNA (ssDNA) of stable stem-loop structures (SLSs). ssDNA is inherently fragile and also an optimal target for AID. We speculate that GCV could have been initiated by the targeting of hypermutable bases in ssDNA state in stable SLSs, plausibly by AID. We have observed that the frequency of GCV-like events is significantly higher in rearranged IGHV3-23-(*)01 sequences from healthy individuals compared to that of CVID patients. We did not observe GCV-like events in rearranged IGHV3-23-(*)01 sequences from AID-deficient patients. GCV, unlike somatic hypermutation (SHM), can result in multiple base substitutions that can alter many amino acids. The extensive changes in antibody affinity by GCV-like events would be instrumental in protecting humans against pathogens that diversify their genome by antigenic shift.

Aborted germinal center reactions and B cell memory by follicular T cells specific for a B cell receptor V region peptide

A fundamental problem in immunoregulation is how CD4(+) T cells react to immunogenic peptides derived from the V region of the BCR that are created by somatic mechanisms, presented in MHC II, and amplified to abundance by B cell clonal expansion during immunity. BCR neo Ags open a potentially dangerous avenue of T cell help in violation of the principle of linked Ag recognition. To analyze this issue, we developed a murine adoptive transfer model using paired donor B cells and CD4 T cells specific for a BCR-derived peptide. BCR peptide-specific T cells aborted ongoing germinal center reactions and impeded the secondary immune response. Instead, they induced the B cells to differentiate into short-lived extrafollicular plasmablasts that secreted modest quantities of Ig. These results uncover an immunoregulatory process that restricts the memory pathway to B cells that communicate with CD4 T cells via exogenous foreign Ag.

Conceptual bases for quantifying the role of the environment on gene evolution: the participation of positive selection and neutral evolution

To demonstrate that a given change in the environment has contributed to the emergence of a given genotypic and phenotypic shift during the course of evolution, one should ask to what extent such shifts would have occurred without environmental change. Of course, such tests are rarely practical but phenotypic novelties can still be correlated to genomic shifts in response to environmental changes if enough information is available. We surveyed and re-evaluated the published data in order to estimate the role of environmental changes on the course of species and genomic evolution. Only a few published examples clearly demonstrate a causal link between a given environmental change and the fixation of a genomic variant resulting in functional modification (gain, loss or alteration of function). Many others suggested a link between a given phenotypic shift and a given environmental change but failed to identify the underlying genomic determinant(s) and/or the associated functional consequence(s). The proportion of genotypic and phenotypic variation that is fixed concomitantly with environmental changes is often considered adaptive and hence, the result of positive selection, even though alternative causes, such as genetic drift, are rarely investigated. Therefore, the second aim herein is to review evidence for the mechanisms leading to fixation.

Somatic translocation and differential expression of Ig mu transgene copies implicate a role for the Igh locus in memory B cell development

Memory B cells of mice with Ig mu transgenes often carry transgene copies that have moved into the Igh locus via somatic translocation. This phenomenon has been attributed to a selection pressure for somatic hypermutations, which generally are observed at much higher frequencies in translocated copies than in ectopic copies. We tested this idea by immunizing Ig-mu transgenic mice in a manner designed to select B cells that required only one V(H) mutation for a switch in antigenic specificity and recruitment into the memory pool. Despite the minimal mutation requirement, hybridomas carrying somatic translocations to the Igh locus were obtained. Importantly, this occurred despite the fact that translocated and untranslocated mu-transgenes were mutated comparably. Evidently, a strong selection advantage was conferred upon B cells by the somatic translocations. Among the hybridomas, translocated mu-transgenes were active, while ectopic mu-transgenes were uniformly silent. The translocated copy that had conferred an affinity-based selection advantage was expressed at the highest level. Moreover, translocated copies were differentially expressed among hybridoma members, which belonged to a common post-mutational lineage. This suggests that adjustments in transgene expression levels had occurred during memory cell development. These results indicate that, apart from their potential influences on somatic hypermutagenesis and class switch recombination, elements in the Igh locus promote the selection of memory B cells in another way, possibly by regulating the level of Ig expression at various stages of antigen-driven differentiation.

Evolution of Ig DNA sequence to target specific base positions within codons for somatic hypermutation

Ig variable (V) region genes are subjected to a somatic hypermutation process as B lymphocytes participate in immune reactions to protein Ags. Although little is known regarding the mechanism of mutagenesis, a consistent hierarchy of trinucleotide target preferences is evident. Analysis of trinucleotide regional distributions predicted and we now empirically confirm the surprising finding that the framework 2 region of kappa V region genes is highly mutable despite its importance to the structural integrity and function of the Ab molecule. Interestingly, much of this mutability appears to be focused on the third codon position where synonymous substitutions are most likely to occur. We also observed a trend for high predicted mutability for codon positions 1 and 2 in complementarity-determining regions. Consequently, amino acid replacements should occur at a higher rate in complementarity-determining regions than in framework regions due to the distribution and subsequent targeting of microsequences by the mutation mechanism. Our results reveal a subtle tier of V region gene evolution in which DNA sequence has been molded to direct mutations to specific base positions within codons in a manner that minimizes damage and maximizes the benefits of the somatic hypermutation process.

Gene conversion-like sequence transfers in a mouse antibody transgene: antigen selection allows sensitive detection of V region interactions based on homology

Gene conversion is important for antibody diversification in chickens, rabbits and cows. In mice, however, conversion events appear to be infrequent among endogenous antibody genes. DNA sequence transfer events that resemble gene conversions have been reported for a mouse H chain transgene (VVC(mu)) that contains two closely spaced homologous VDJ segments. Surprisingly, these reported VVC(mu) sequence transfers were found frequently among mouse B cells responding to immunization. Transgene sequence transfers could be occurring at high frequency in responding VVC(mu) B cells or could be occurring at lower frequency with subsequent amplification by preferential antigen selection. To distinguish these possibilities, we have analyzed a second transgene (InVVC(mu)) that is identical to VVC(mu) except that the two VDJ regions have been exchanged in position. We find that transgene sequence transfers are much less frequent among responding B cells in InVVC(mu) mice, demonstrating the importance of selection in the frequent transgene conversions observed in VVC(mu) mice. These results suggest that mice, like other species, can use gene conversion to diversify antibodies. Such diversification events are apparently infrequent, however, and might only be detected among endogenous Ig genes with a favorable arrangement of V genes and an antigenic stimulation that selects cells with conversions. For both VVC(mu) and InVVC(mu) mice, conversion-like sequence transfers are strongly correlated with somatic hypermutation. Based on these results, we hypothesize that, in mice, gene conversions represent infrequent alternative reactions of a homology-based DNA repair process that is central in the somatic hypermutational mechanism.

Somatic Mutation in the Neonatal Mouse

Several mechanisms that diversify the adult immune repertoire, such as terminal deoxynucleotidyl transferase-dependent N region addition, are not available to the neonatal mouse. One important process that contributes to protective immunity in the adult is somatic mutation, which plays a major role in the generation of high affinity memory B cells. It is not clear whether B cells in the neonatal mouse can activate the somatic mutation machinery. To investigate this, we immunized neonates with poly(l-Tyr,l-Glu)-poly-d, l-Ala–poly-l-Lys complexed with methylated BSA, or (4-hydroxy-3-nitrophenyl)acetyl coupled to chicken γ-globulin. Eight to fourteen days after priming, V(D)J rearrangements of known VH genes (VHSM7 family) were screened for mutations using a temperature-melt hybridization assay and oligonucleotide probes specific for complementarity-determining regions I and II; possible mutations were confirmed by sequence analysis. More mutations per sequence were found in heavy chains from neonates immunized with (4-hydroxy-3-nitrophenyl)acetyl coupled to chicken γ-globulin than in those from neonates immunized with poly(l-Tyr, l-Glu)-poly-d,l-Ala-poly-l-Lys complexed with methylated BSA. Mutations were found in heavy chains lacking N regions, suggesting that B cells of the putative fetal lineage can somatically mutate and diversify an initially limited repertoire. Since neonates immunized as early as 1 or 2 days after birth had mutations, the somatic mutation machinery can be activated soon after birth, suggesting that early vaccination should result in affinity maturation and protective immunity in the neonate.

Somatic origin of T-cell epitopes within antibody variable regions: significance to monoclonal therapy and genesis of systemic autoimmune disease

During an immune response, specific antibody variable region genes are diversified by a somatic point mutation process that generates de novo "foreign" V-region sequences. This creates an interesting problem in immune regulation because B cells are highly proficient at self-presenting V-region peptides in the context of class II MHC. Though our studies indicate that the corresponding T-cell repertoire attains a state of tolerance to germline-encoded antibody V-region diversity, it is presently unknown whether the same is true of mutationally generated diversity. On the basis of immunoregulatory considerations, we hypothesize that contact exclusion or tolerance normally precludes T cells from helping B cells via self-presented mutant V-region peptides. The lack of recurrent somatic mutations that create known T-cell epitopes in antibody V regions lends some support to this idea. In contrast, our studies of spontaneously autoreactive B cells in systemic autoimmune disease strongly suggest that precursors of such cells are recruited by T-cell help directed to self-presented mutant idiopeptides. Failures in tolerance or contact exclusion mechanisms may be responsible for this apparently abnormal event. In addition to their importance in immune regulation, somatic mutations or other differences from germline-encoded V-region sequence may be largely responsible for undesirable patient responses to therapeutic monoclonal antibodies. These reactions might be averted or diminished by inducing tolerance in the T-cell repertoire with synthetic peptide correlates of non-germline-encoded V-region sequences in humanized antibodies.

The molecular basis of somatic hypermutation of immunoglobulin genes

Somatic hypermutation amplifies the variable region repertoire of immunoglobulin genes. Recent experimental evidence has thrown light on various molecular models of somatic hypermutation. A link between somatic hypermutation and transcription coupled DNA repair is shaping up.

Somatic hypermutation

For the generation of secondary response antibodies, immunoglobulin genes are subjected to hypermutation. Cells expressing antibodies with higher affinity are then selected by antigen. Recent clues to the mechanism of hypermutation come from experiments using transgenic mice enabling analysis of the controlling cis-acting elements and the intrinsic features of the hypermutation, dissociated from the effects of antigenic selection.

Analysis of sequence transfers resembling gene conversion in a mouse antibody transgene

The role of gene conversion in murine immunoglobulin gene diversification is unclear. An antibody gene construct designed to provide the homologous donor and acceptor sequences required for conversion mechanisms was produced and used to generate transgenic mice. When these transgenic mice were immunized, DNA sequence transfers between tandem transgene VDJ regions were detectable and resembled gene conversion events. There is a strong link between these conversion-like sequence transfers and transgene somatic hypermutation, suggesting that both processes might occur at the same stage of B cell differentiation.

Mapping the upstream boundary of somatic mutations in rearranged immunoglobulin transgenes and endogenous genes

Mammalian B-cell specific somatic hypermutation contributes to affinity maturation of the antibody response. This mutator activity is highly focused on rearranged immunoglobulin variable regions, but the underlying mechanism remains to be elucidated. In an effort to gain insights into the mechanism of somatic hypermutation, the precise distribution and frequency of mutations upstream of murine immunoglobulin genes was determined by examining the same variable gene segments when mutated in different B-cell lines. Immunoglobulin sequences analysed included kappa light chain transgenes bearing mutated V kappa 24 variable regions, and the endogenous V kappa gene isolated from myeloma MOPC167, which also exhibits mutations in the variable region. In addition, mutated endogenous VH1 gene segments of the S107 heavy chain variable gene family were also examined. For both VH1 and V kappa 24, somatic mutations were generally not found upstream of the leader intron, even in genes which exhibited a high mutation frequency in the variable region itself. The 5' somatic mutation boundary identified in immunoglobulin transgenes overlaps the boundary observed in endogenous genes, suggesting that both share cis-elements required for defining the mutable domain. Furthermore, the location of this 5' boundary appears not to change when these immunoglobulin genes are examined in different cell lines. These data may be indicative of a defined start site for immunoglobulin mutator activity.

Junctional diversification in the generation of the precursor of a discrete immune response

Phosphocholine (PC)-specific antibodies that arise in the mouse in response to Proteus morganii (PM) and use V1-DFL16.1-JH1 are characterized by a number of recurring mutations. Most striking is an invariant A for G substitution in codon 95 of VH which results in an asparagine instead of aspartate at that position. Because of the apparent importance of this substitution in an anti-PC(PM) response, we wanted to determine the molecular basis for this base change. A cDNA library derived from pre-immune splenic B cells was examined for the frequency of VDJ containing the A substitution at 95 and the presence of additional point mutations in these sequences. Six different cDNA were isolated which contained an A substitution at the VD junction (frequency 0.00009); a seventh positive cDNA could not be examined. The V segments of four of these cDNA matched known germline genes and were, therefore, unmutated. Two others closely matched V in families whose members have not all been characterized, hence, it is not known whether the mutations observed are somatic or germline in origin. Sequences of 35 cDNA clones, containing the same V segment but differing in D, J and junctional nucleotides, revealed no mutations. These results indicate that the A substitution generated at codon 95 is most likely a product of V-DJ joining.

Molecular evolution of the human immunoglobulin E response: high incidence of shared mutations and clonal relatedness among epsilon VH5 transcripts from three unrelated patients with atopic dermatitis

We have analyzed the nucleotide sequences of 19 epsilon VH5 transcripts derived from in vivo isotype switched peripheral blood B cells of three patients with atopic dermatitis. Comparison with the patients' own germline VH5 gene segments revealed that the epsilon transcripts were derived from both functional members of the human VH5 gene family and harbored numerous somatic mutations (range 5-36 per VH5 gene). In two patients, we detected clonally related but diverged transcripts, permitting the construction of a genealogical tree in one patient. We observed a high proportion of shared silent (S) and replacement (R) mutations among epsilon VH5 sequences derived from all three individuals, even among transcripts descending from the two different germline VH5 gene segments. A remarkably high number of these mutations is shared with previously reported VH5 genes encoding antibodies with defined specificities. The shared S mutations, and likely a fraction of the R mutations, appear to mark preferential sites ("hot spots") of somatic hypermutations in human VH5 genes. The distribution of R and S mutations over complementarity determining region and framework regions in the majority of VH regions deviated from that characteristic of antigen-driven immune response. We hypothesize that the V regions of immunoglobulin E-bearing B cells have accumulated "selectively neutral" mutations over extended periods of clonal expansion, resulting in unusual R/S ratios. We propose that the molecular characteristics of the epsilon VH regions in atopic dermatitis may be representative of antigens that recurrently or chronically stimulate the immune system.

Oligonucleotide-directed mutagenesis of antibody combining sites

We review here our attempts to achieve a better understanding of the structure--function relationship of antibody combining sites, and to gain insights into the engineering of antibodies with desired specificity and affinity. We have focused on a model system--antibodies to the hapten p-azophenylarsonate (Ars) derived from A/J mice. Oligonucleotide-directed mutagenesis was used to alter the sequence of the variable region genes of such anti-Ars antibodies. Mutant antibodies were generated in hybridoma cells following transfection of the altered genes, and the effects of the primary structure changes on antibody specificity, affinity, and idiotypic expression were assessed. These studies suggest that an antibody combining site with basic specificity for an antigen could be created by introducing a set of a few amino acid residues in the complementarity determining regions, and that the affinity of such a site could be improved one substitution at a time in a sequential manner.

Germ line variable regions that match hypermutated sequences in genes encoding murine anti-hapten antibodies

We asked whether there are germ line immunoglobulin variable (V) segments that match sites of hypermutation in V regions encoding murine antibodies. Murine germ line DNA was probed with a panel of short deoxyoligonucleotides identical in sequence to segments of hypermutated V regions from hybridomas generated in the BALB/c response to the hapten 2-phenyloxazolone (Ox). Germ line sequences that match mutations in both heavy and kappa light chain V regions were identified, and clones of some of these germ line V segments were obtained. Comparison of these clones with hypermutated V regions revealed regions of identity ranging in size from 7 to over 50 nucleotides. In an effort to separate the effects of antigen selection from the mutagenic process, we also searched for matches to a panel of silent mutations in VH regions from germinal center B cells. Fourteen silent mutations occur among a collection of 36 hypermutated VH regions from two separate germinal centers of C57BL/6 mice stimulated with the hapten 4-hydroxy-3-nitrophenyl. Matches to nine of these silent mutations can be found among published sequences of C57BL/6 VH regions of the J558 family. Taken together, these data are consistent with the possibility that a template-dependent mutational process, like gene conversion, may contribute to somatic hypermutation.

Non-random features of the repertoire expressed by the members of one V kappa gene family and of the V-J recombination

The 5' and 3' flanking sequences of 14 members of the V kappa Ox (VK 4/5) gene family of BALB/c mice have been established. The family was unusual in the number of bases between the codon for Pro 95 and the heptamer sequence; most members contained four but there were also examples of none. A conserved leader sequence was used to amplify the genomic DNA of rearranged genes in order to analyze the spleen B cell repertoire of non-immunized animals. The library contained many members with virtually identical sequences to one or other of the already known members of the family. In addition, there were repeats of other sequences, allowing the definition of 12 hitherto undefined members of the family. Only 3 out of 96 could have originated by gene conversion, or as artefacts of the amplification procedure, and only 2 were putative somatic mutants. The frequency of expression of different members of the V kappa Ox gene family was not random, and some germ-line genes were unrepresented in the library. The high frequency of V kappa Ox1-J kappa 5 is in line with the dominance of this combination in the oxazolone response. An analysis of the junctional segment showed that although in most cases the diversity was due to trimming, there were exceptions indicating de novo additions (N or P bases). The average number of bases trimmed from the V kappa and the J kappa segments was not the same. There was no correlation in the number of bases trimmed from V kappa or J kappa in each recombination. The implications of asymmetric trimming in terms of the mechanism of recombination are discussed.

B-cell proliferation initiated by Ia cross-linking and sustained by interleukins leads to class switching but not somatic mutation in vitro

Somatic mutations that are acquired by antibody V genes of antigen-stimulated B cells ultimately provide the clonal diversity from which memory B cells are selected during immune responses to T-cell-dependent antigens. Somatic mutations apparently are not acquired when B cells are stimulated by mitogens nor when they participate in immune responses to T-cell-independent antigens. Since the basis of T-cell-dependent humoral immunity is T-cell recognition of processed antigen in the context of class II major histocompatibility glycoproteins (Ia) on the B-cell surface, we sought to determine whether the ligation of Ia on B cells induces somatic mutation. B cells were stimulated in vitro by a procedure in which their proliferation was dependent upon ligation of surface Ia with antibody. Sequences of hybridoma V genes derived from these B cells revealed no somatic mutations despite prolonged stimulation in vitro and the induction of immunoglobulin secretion and switching to isotypes characteristic of T cell-dependent humoral immunity. We infer that Ia-mediated signalling and isotype switching are not causally related to somatic mutation. The avenue of differentiation that leads to somatic mutation in memory B cells is apparently separable from that leading to proliferation, immunoglobulin secretion and switching.

A new germline VH36-60 gene is used in the neonatal primary and adult memory response to (T,G)-A--L

Our previous studies of the neonatal primary response to (T,G)-A--L showed that the majority of anti-(T,G)-A--L antibodies bind the copolymer L-Glu:L-Tyr (GT), share idiotypy (Id), and use the H10 germline VH gene from the VHJ558 family and a V kappa 1 gene. We also identified two hybridomas from different neonatal donors that produced GT+, Id+ antibodies using a V kappa 1 gene with a VH gene from the VH36-60 family. In the study reported here, we show that both neonatal hybridomas use the same germline VH gene from the VH36-60 gene family. However, the VH gene sequence is different from previously identified germline genes of the VH36-60 gene family. To determine whether the expressed heavy chain gene had undergone somatic mutation, we isolated the corresponding germline gene from kidney DNA. Sequence analysis of this gene shows that it is a new member of the VH36-60 family which is not mutated in the neonatal antibodies. Furthermore, the deduced amino acid sequences of the two neonatal antibodies are identical not only in the VH region but also in the VH-D-JH joins, suggesting that there is a strong selection for CDRIII among neonatal anti-(T,G)-A--L antibodies using this germline gene (designated here as VH3A1) with a V kappa 1 gene. Also, the VH gene from the VH36-60 family that we showed previously was used by an adult memory B cell clone specific for (T,G)-A--L, can now be identified as a rearrangement of the VH3A1 germline gene. Elucidation of the germline variable region genes that are used in the antigen-specific neonatal response will help us understand the mechanisms that shape the preimmune B cell repertoire during B cell development.

Different epitope structures select distinct mutant forms of an antibody variable region for expression during the immune response

Antibody variable (V) regions that initially differ from one another by only single amino acid residues at VH-D and D-JH segment junctions (termed canonical V regions) can be elicited in strain A/J mice by three different haptens. Among such V regions an amino acid substitution due to somatic mutation is recurrently observed at VH CDR2 position 58, regardless of which of these haptens is used for immunization. This substitution confers upon a canonical V region a generic increase in affinity for all the haptens. Conversely, the type of amino acid substitution at VH position 59 resulting from somatic mutation that is recurrently observed among such V regions changes with the eliciting hapten, in a manner that correlates directly with the cognate affinity increases (or decreases) for hapten conferred by the observed substitutions. This small subregion of VH CDR2 therefore plays a major role in determining both affinity and specificity for antigen. The data confirm that affinity for antigen is of pivotal importance in determining the degree of selection of different mutant forms of a V region. Moreover, during an immune response a sufficiently diverse mutant repertoire can be generated from a single canonical V region to allow adaptation to increase affinity for three different epitopes.

Boundaries of somatic mutation in rearranged immunoglobulin genes: 5' boundary is near the promoter, and 3' boundary is approximately 1 kb from V(D)J gene

To investigate why somatic mutations are spatially restricted to a region around the rearranged V(D)J immunoglobulin gene, we compared the distribution of mutations flanking murine V gene segments that had rearranged next to either proximal or distal J gene segments. 124 nucleotide substitutions, nine deletions, and two insertions were identified in 32,481 bp of DNA flanking the coding regions from 17 heavy and kappa light chain genes. Most of the mutations occurred within a 2-kb region centered around the V(D)J gene, regardless of which J gene segment was used, suggesting that the structural information for mutation is located in sequences around and within the V(D)J gene, and not in sequences downstream of the J gene segments. The majority of mutations were found within 300 bp of DNA flanking the 5' side of the V(D)J gene and 850 bp flanking the 3' side at a frequency of 0.8%, which was similar to the frequency in the coding region. The frequency of flanking mutations decreased as a function of distance from the gene. There was no evidence for hot spots in that every mutation was unique and occurred at a different position. No mutations were found upstream of the promoter region, suggesting that the promoter delimits a 5' boundary, which provides strong evidence that transcription is necessary to generate mutation. The 3' boundary was approximately 1 kb from the V(D)J gene and was not associated with a DNA sequence motif. Occasional mutations were located in the nuclear matrix association and enhancer regions. The pattern of substitutions suggests that there is discrimination between the two DNA strands during mutation, in that the four bases were mutated with different frequencies on each strand. The high frequency of mutations in the 3' flanking region and the uniqueness of each mutation argues against templated gene conversion as a mechanism for generating somatic diversity in murine V(D)J genes. Rather, the data support a model for random point mutations where the mechanism is linked to the transcriptional state of the gene.