Sequences of the VH and VL regions of murine monoclonal antibodies against 3-fucosyllactosamine

Many mAb that bind the carbohydrate antigenic determinant 3-fucosyl-lactosamine (3-FL), Gal beta 1-4[Fuc alpha-3]GlcNAc-R have been raised in BALB/c mice, and we are studying the structure and regulation of these antibodies. In this report, we present the first information about their amino acid sequences and the Ig gene segments used to encode them. V regions of the H and L chains of three anti-3-FL antibodies, PMN6, PMN29, and PM81, were sequenced by a combination of mRNA and amino acid sequencing. The L chain sequences of PMN6 and PM81 antibodies indicate that their VK and JK regions are encoded by VK24B and JK1 germ-line genes, respectively. The nucleotide and amino acid sequences of the H chains suggest that the three anti-3-FL antibodies are encoded by the VH441 gene segment of the X24 VH family, and this conclusion was supported by Southern filter hybridization with VH441 and JH3-JH4 probes. PMN29 has at least 11 amino acid substitutions, which is an unusually large amount of somatic mutation for an IgM antibody. Previous analyses of BALB/c genomic libraries with VHX24 and VH441 probes make it unlikely that this VH family contains additional germ-line genes, but this possibility cannot be excluded. All three antibodies use the DQ52 and JH4 gene segments. The single VH and VL gene segments used to encode the anti-3-FL antibodies is in contrast to the multiple VH and VL segments used by antibodies against other carbohydrate Ag such as alpha 1-6 dextran and group A streptococcal carbohydrate. VH441 also encodes the VH regions of antibodies against galactan and levan (beta 2-6 fructosan). The similarities among VH segments of antibodies against 3-FL, levan, and galactan, and the striking differences in their CDR3 sequences, suggest that CDR3 plays an important role in the formation of the Ag binding site. The use of a single VH segment from the smallest VH gene family by antibodies against at least three different carbohydrate determinants is noteworthy. It raises the possibility that the amino acid sequence encoded by VH441 has some general structural features that make it particularly well adapted for binding to carbohydrate sequences.     

V region sequences of anti-DNA and anti-RNA autoantibodies from NZB/NZW F1 mice

The V region sequences of two anti-DNA (A52, D42) and two anti-RNA (D44, D444) autoantibodies, derived from lupus prone NZB/NZW F1 female mice, were determined by mRNA sequencing. The sequences had the following features: 1) there was no clear sequence relationship between anti-DNA and anti-RNA antibodies; 2) there were no major similarities between any of the L chain sequences and each VL gene segment belonged to a different mouse VK subgroup; 3) the H chains of the two anti-RNA antibodies showed closely related sequences of VH gene segments and very similar third complementarity determining regions (CDR3); 4) the H chains of the two anti-DNA antibodies had VH segments belonging to different VH gene families but had a unique and similar combination of D segments and junctional sequences, suggesting a common recognition element for Ag and/or for idiotypic regulation in the H chain CDR3; and 5) the VH gene segment of one anti-DNA antibody (D42) was found to be very similar to the VH gene segment of a CBA mouse hybridoma antibody (6G6) which binds to the environmental Ag phosphocholine. The three-dimensional structure of the Fv-region of the anti-DNA antibody (D42) was modeled by computer and a stretch of poly(dT), ssDNA was docked to a cleft in the antibody combining site, formed by the three H chain CDR and by CDR1 and CDR3 of the L chain. The cleft is characterized by a preponderance of arginine and tyrosine residues, lining both the walls and base of the cleft.     

Molecular basis of an isogeneic anti-idiotypic response.

The nucleotide sequences of the variable region genes expressed in the heavy and light chains of six isogeneic anti-idiotope antibodies recognizing idiotopes on two closely related antibodies with specificity for the hapten (4-hydroxy-3-nitrophenyl)acetyl (NP) were determined. In two independently derived anti-idiotope cell lines the same or strongly homologous V kappa, VH and D region genes had originally been rearranged. The two lines express long and partly homologous N sequences (presumed to be not of germ line origin) at the border of D, resulting in CDR3s of unusual length. An unusually long CDR3, partly encoded by N sequences, is also present in the heavy chain of a third anti-idiotope antibody. The VH regions of the three remaining anti-idiotope antibodies originate from a single VH gene which belongs to the same VH group as the VH genes expressed in the other anti-idiotopes. Two of these antibodies, expressing similar V, D and J elements, had been isolated from the same mouse and appear to have diverged from the same B cell precursor by at least two rounds of somatic mutation. Somatic point mutations have occurred in most, if not all anti-idiotope V region sequences. In two instances somatic mutations in J increase the structural homology between anti-idiotopes. The anti-idiotypic response in this system is thus genetically restricted and may depend upon the selection of non-germ line sequences, suggesting an explanation for the low frequency at which anti-idiotope antibodies are expressed in this system.     

VDJ genes in VHa2 allotype-suppressed rabbits. Limited germline VH gene usage and accumulation of somatic mutations in D regions

In this study we investigate the molecular genetic basis for VHa- Ig. Knowing that the expression of VHa allotype Ig is suppressed by neonatal injection of rabbits with anti-VHa allotype antibody, and that the decreased level of VHa allotype Ig, VHa+, in the suppressed rabbits is compensated for by an increase in VHa- Ig, we determined the nucleotide sequences of 41 VDJ genes from a2/a2 rabbits neonatally suppressed for the expression of a2 Ig. We compared these nucleotide sequences to each other and identified two groups of VH sequences. We predict that the molecules of each group are encoded by one germline VH gene. Inasmuch as VHa+ Ig is encoded predominantly by one germline VH gene, VH1, it appears that more than 95% of the VDJ repertoire of rabbits may be encoded by as few as three germline VH genes. A genomic VDJ gene whose VH sequence was similar to those of group I molecules was expressed in vitro and was shown by ELISA to encode molecules of the VHa- allotype, y33. Analysis of the D regions in the VDJ gene indicated that germline D2b and D3 gene segments were preferentially used in the VDJ gene rearrangement. A comparison of sequences of D regions of the 41 VDJ gene rearrangements in 3-, 6-, and 9-wk-old rabbits to sequences of germline D gene segments showed an accumulation of mutations in the D region. Inasmuch as we have previously shown that V regions of rabbit VDJ genes are diversified, in part, by somatic gene conversion, it appears now that rabbit VDJ genes diversify by a combination of somatic mutation and somatic gene conversion.     

Patterns of mutations and selection in antibodies to the phosphocholine-specific determinant in Proteus morganii

The contribution of somatic mutation to the generation of an antibody response was investigated by using the phosphocholine (PC) determinant in the bacterium Proteus morganii as the model Ag. The response to this determinant is restricted to a single VH/VL pair and apparently is derived from only one or two precursors per mouse. In this study we examined hybridoma antibodies from nine individual mice which produced representatives of 12 different clones. We found that all antibodies reactive with the PC Ag of P. morganii contained somatic mutations; the number ranged from 2 to 20. Two clusters of mutations were observed, one in complementarity-determining residue (CDR) 2 and the other in CDR 3 of VH. Examination of a three-dimensional model of M603, an antibody with the same V region composition as the anti-PC antibodies under study, showed that these clusters occupied an area of the binding site which presumably interacts with carrier elements of the PC epitope in P. morganii. A high incidence of recurring mutations were found in both clusters, and one of these was invariant, leading to an Asn for Asp substitution at 95. Ag binding studies with these antibodies and an additional one, which was unmutated except for the invariant substitution at 95, showed that: 1) antibodies having only the 95Asn mutation failed to bind the PC Ag of P. morganii, 2) addition of a second recurring mutation, at 52a (CDR 2), was sufficient to create strong binding to the P. morganii Ag, and 3) accumulation of mutations was directly correlated with increased binding activity for Ag. These results show that somatic mutations play a critical, if not essential, role in generating specificity for this PC Ag, and that Ag, and most likely a carrier element of the epitope, is a primary force in the continued selection and expansion of Ag-reactive B cells.     

Germ-line affinity and germ-line variable-region genes in the B cell response

The predominance of germ-line genes in IgM expression was evaluated from the nucleotide sequences of mRNA, derived from 10 hybridoma cell lines, coding for the VH and VL regions of anti-5-dimethylaminonaphthalene-1-sulfonyl (anti-Dns) IgM antibody. At least six germ-line VH gene segments distributed among four families are used in this response. Seven of the 10 independently rear-ranged VH genes were identified as germ line, with the other three possibly germ line. In all of them the D and JH portions retained the germ-line sequences of the D and JH segments from which they were derived. Maximum diversity was found in the D segments and the use of noncoded nucleotides at the VH-D and D-JH junctions. Of the eight cell lines expressing the lambda light chains, all were germ line and involved the three subtypes. Maximum affinity for the homologous ligand was found among the seven cell lines identified as expressing germ-line gene segments. Thus any somatic mutation among the remaining 3 cell lines did not provide enhanced affinity and the observed affinity of each cell line can be described as germ-line affinity. It is further suggested that the anti-Dns selectivity of the IgM antibodies is associated primarily with the CDR3 regions.     

Study of idiotopic suppression induced by anti-cross-reactive idiotype monoclonal antibody in the anti-p-azophenylarsonate antibody response

A/J mice immunized with p-azophenylarsonate coupled to keyhole limpet hemocyanin produce antibodies expressing a cross-reactive idiotype (CRIA). The pretreatment of A/J mice with anti-idiotypic polyclonal or monoclonal antibody directed against the major cross-reactive idiotype (CRIA) borne by p-azophenylarsonate-specific antibody can lead to idiotypic suppression. In this study, we investigate this idiotypic suppression by using four mAb2 (E4, H8, E3, 2D3) recognizing distinct idiotopes whose expression is related to the presence of particular gene segments of the heavy chain V region. 2D3 expression has been related to the presence of some amino acid in the CDR2 region of the VH gene segment derived from the germ line VH IdCR11. So far, the latter is the only germ-line gene coding for CRIA+ antibody that has been identified in the A/J genome. E4 and H8 expression has been related to the use of a particular D segment, whereas E3 expression has been attributed to certain combinations of D and JH segments. Therefore, we might expect independent regulation of the expression of those various idiotopes in relation to the mechanism of gene recombination. Indeed, we observed that 2D3-suppressed A/J mice still produce the three other idiotopes, suggesting the recombination of those particular D and J segments with a different VH gene. Such a gene has been identified in the genome of BALB/c mice. A/J mice pretreated with one of the other three mAb2 are generally cosuppressed for the expression of E4, H8, and E3, but they still produce 2D3+ antibody. In this case, the IdCR11 VH germ-line gene is most probably recombined with different D and J segments. Molecular evidence for the existence of such molecules has also been presented in the literature. So our serologic data on idiotopic suppression in the arsonate system can be compared with recent data provided by molecular genetics.     

The BALB/c secondary response to the Sb site of influenza virus hemagglutinin. Nonrandom silent mutation and unequal numbers of VH and Vk mutations

We have determined the nucleotide sequences of the expressed VH and Vk genes from 13 secondary (2 degrees) hemagglutinin (HA) (Sb) specific hybridomas derived from a single mouse. These antibodies share an Id, H37-68 (68Id) that dominates the 2 degrees HA(Sb) response in this mouse, but is rare or absent from 2 degrees responses of other mice. We find that these antibodies derive from five clones. The H chains of these antibodies are encoded by a single VH gene joined to a variety of DH and JH genes. The length of complementarity-determining region (CDR) 3 and sequence of the D-J junction are restricted, suggesting selection on CDR3 of the H chain. The L chains are more diverse. In the presented examples, they are encoded by the Vk21C and Vk21E genes and a Vk9 gene, and are joined to Jk1, 2, or 4. Each antibody is extensively mutated. The nature and distribution of the mutations suggests that 68Id-producing cells have been selected by Ag, although there are differences regarding the domain (VH, Vk, or both) in which mutations were selected. The implications of these findings on the idiosyncratic nature of the 68Id antibody response to HA(Sb) are discussed. There are two unusual characteristics regarding somatic mutation in these hybridomas. Whereas the expressed VH and Vk21 genes appear to have accumulated mutations at a high rate (1 to 1.5 x 10(-3)/base pairs/division, the expressed Vk9 genes appear to have accumulated mutations at a 5 to 15-fold lower rate than the expressed VH genes in the same cells. There is also a surprisingly high number of parallel silent somatic mutations in the VH genes, of which all but one are clustered to a 28-bp region in framework region 2 and CDR 2-encoding segments. The probability that this could have occurred by a random mutational process is essentially zero.     

Germline V genes encode viable motheaten mouse autoantibodies against thymocytes and red blood cells

Autoantibodies against thymocytes and RBC may contribute to the pathophysiology of homozygous viable motheaten (mev) autoimmune disease. Whether the production of these autoantibodies in mev mouse results from polyclonal nonspecific B cell activation or specific Ag-driven stimulation is not known. To understand the mechanisms involved in the induction of antithymocyte autoantibody response in mev mouse, we have studied the fine antigenic specificity, structure, and origin of three antithymocyte autoantibodies derived from mev splenic B cell hybridomas. Western blot analysis showed that these mAb bind to polypeptides of 33 and 105 kDa present in RBC and thymocytes, respectively. Additional specificities for the epitopes present in other polypeptides distinguished these three autoantibodies. Northern hybridization and flow microfluorimetry analysis indicated that these hybridomas are derived from the Ly1+ B cell subset. These autoreactive Ly-1 B cell hybridomas, chosen on the basis of their specificity, expressed L chain V genes from a single VK family (VK9) and VH genes from J606 and S107 families. Hybridomas UN34.11 and UN42.5 expressed the VK9 gene identical to that used by peritoneal Ly1+ B cells from various mouse strains and malignant B lymphoma cells secreting anti-mouse RBC treated with proteolytic enzyme bromelin and anti-SRBC antibodies. The third hybridoma, S2-14.2, used a VK9 gene identical to that expressed by MOPC41. None of the VK genes encoding these autoantibodies showed any somatic mutations. In the case of VH genes, the two hybridomas UN42.5 and S2-14.2 derived from two separate fusions, used identical VH genes from the J606 family. The third hybridoma UN34.11 used unmutated V11 germline VH gene, a member of the S107 family. Southern hybridizations, using oligonucleotide probes specific for CDR1 and CDR2, showed that the VH genes encoding the J606 autoantibodies were derived from a germline gene found in the 6.7-kb fragment of EcoRI-digested germline DNA. This germline VH gene is distinct from VH22.1 germline gene that codes for antigalactan antibodies. Sequence analysis of this gene showed perfect homology with the rearranged VH genes confirming the lack of somatic mutations. Thus, our data demonstrate that antithymocyte antibody response occurring in mev mouse is polyclonal and it involves Ly-1 B cells expressing unmutated germline VH and VK genes. These results indicate that antigen driven stimulation may not play an important role in the induction of anti-thymocyte antibody response in mev mouse.     

Characteristics of murine monoclonal anti-CD4. Epitope recognition, idiotype expression, and variable region gene sequence.

We have characterized a series of mouse monoclonal anti-CD4 and describe both their CD4 epitope recognition and Id expression. We also determined the V region gene sequences of these antibodies in an attempt to correlate epitope recognition and Id expression with V region sequence. All of these preparations recognize epitopes that cluster around the HIV gp120 binding site on the human CD4 molecule. However, we observed differences in epitope recognition among the anti-CD4 preparations, based on either competitive inhibition assays or functional assays, such as syncytium inhibition. Analysis of Id specificities using a polyclonal anti-Id generated against anti-Leu 3a indicated that five of the seven monoclonal anti-CD4 expressed a shared Id. Based on V region gene sequences, the V region kappa-chain (V[kappa]) from each of the seven antibodies was encoded by the V[kappa]21 gene family and expressed the J[kappa]4 gene segment. Those preparations that expressed the shared Id with anti-Leu 3a have virtually identical V[kappa] sequences, with a high degree of homology in the CDR. The VH region gene sequences of six of the seven antibodies also shared overall homology and appeared to be encoded by the J558 VH gene family. The seventh anti-CD4 VH region is encoded for by the VHGAM gene family. The majority of these antibodies used JH3 gene segment, although the JH2 and JH4 gene segments were also represented. In addition, several of these antibodies share a common sequence organization within their V-D-J joining regions that appears to involve N and P sequences to generate unique D segments. Together, these data suggest that differences in epitope recognition among the monoclonal anti-CD4 may reflect sequence variability primarily within the CDR3 region of both V[kappa] and VH. The basis for the detection of a shared Id most likely reflects the high degree of homology within the V[kappa] region sequences. In addition, these data, which are based on a limited analysis, suggest the possible restricted use of V region germ-line gene families in the secondary antibody response of BALB/c mice to specific epitopes on the human CD4 molecule.     

The specificity properties that distinguish members of a set of homologous anti-digoxin antibodies are controlled by H chain mutations

Five murine A/J strain anti-digoxin mAb (35-20, 40-40, 40-120, 40-140, and 40-160) have highly homologous H and L chain V regions, only differing by somatic mutation, yet differ in affinity and specificity. The availability of the VH and VL genomic clones from one hybridoma, 40-140, has now allowed studies involving in vitro mutagenesis and chain recombination among these five hybridomas. To determine the relative contributions of the mutations found in either VH or VL to the overall binding properties of these antibodies, we recombined the 40-140VH with the VL of each hybridoma. The 40-140VH gene was transfected into hybridoma variants that produce only VL. The recombinant antibodies show that the mutations present in VH, rather than in VL, affect the fine specificity properties of these antibodies, whereas, the mutations among both VH and VL chains are important in determining antigen affinity. From mutations present in VH that affect fine specificity properties, the comparison of the antibody sequences, and from the previously measured binding properties, we predicted and tested selected VH mutations for their ability to alter specificity or affinity by doing site-directed in vitro mutagenesis. The results for the somatic mutations found in this group of antibodies show: 1) VH mutations control the fine specificity properties that distinguish different members of this group; 2) in particular, VH residues 54 and 55 in CDR2 control the distinguishing characteristics of specificities between these antibodies; and 3) by mutagenesis, we had the unusual result of being able to alter Ag specificity without affecting affinity. A computer model of the 40-140 antibody binding site was generated which indicates that VH residues 54 and 55 are highly accessible.     

VH restriction among human cold agglutinins. The VH4-21 gene segment is required to encode anti-I and anti-i specificities.

We previously reported that human autoantibodies with cold agglutinin activity contained a single human VH gene segment (VH4-21) which was also responsible for the cross-idiotypic specificity characteristic of the cold agglutinin response. To confirm and extend this observation we have analyzed at the nucleotide level the H and L chains of six new cold agglutinin molecules derived from different patients. We found that regardless of whether the antibody recognizes the i or the I red cell Ag, restriction at the VH gene segment level is absolute. We also found that even in the absence of somatic mutation the VH4-21 gene segment can encode both anti-i and anti-I specificities. Finally, although the VH4-21 gene segment is essential for cold agglutinin activity, the other genetic elements that contribute to the V region of the antibody molecules can be extremely diverse. The structural information provided in this report sharply restricts the requirement for encoding pathogenic cold agglutinin activity to one of the components of the H chain V region, specifically the VH gene segment. The implications of this apparently absolute requirement for a single VH gene segment, unprecedented in the human autoimmune response, are discussed.     

Nucleotide sequence of a human monoclonal anti-idiotypic antibody specific for a rabies virus-neutralizing monoclonal idiotypic antibody reveals extensive somatic variability suggestive of an antigen-driven immune response

We previously described the isolation and characterization of a human monoclonal anti-idiotypic antibody (Ab2) isolated from EBV-transformed human PBL after immunization with rabies vaccine. The present study concerns the molecular characteristics of the Ab2 and the germ-line elements that gave rise to it. The H chain of this antibody derives from the small VHV family of human V region gene segments. Parallel studies on the germ-line VHV gene isolated from the same individual revealed that the expressed molecule contains 19 nucleotide differences in the VH gene segment. The D segment of Ab2 could have arisen by a D to D fusion; the J segment is a JH6. Extensive somatic variation evident in the H chain variable region of this naturally arising monoclonal anti-idiotypic antibody suggests that this Ab2, the product of a CD5+ B cells, was the consequence of an Ag-driven immune response.     

Two families of monoclonal antibodies to alpha(1----6)dextran, VH19.1.2 and VH9.14.7, show distinct patterns of J kappa and JH minigene usage and amino acid substitutions in CDR3.

Nine groove-type mAb to alpha(1----6)dextran were cloned and sequenced. Together with previous reports from this laboratory, the VH and VL of 34 mAb have been sequenced, in which 10 VH19.1.2 and 11 VH9.14.7 combined with the V kappa-Ox1 gene to form two major families of anti-alpha(1----6)dextrans. The same D minigene (DFL16) was used by all VH19.1.2 and VH9.14.7 mAb; however, the patterns of JH and J kappa usage are quite different. VH19.1.2 mAb used only JH3 and J kappa 2, whereas VH9.14.7 mAb used three JH (JH1, JH2, and JH3) and all four active J kappa (J kappa 1, J kappa 2, J kappa 4, and J kappa 5). Relative uniformity in the lengths of VH CDR3 and the junctional sequences is seen in both families. Some mAb from different mouse strains share common structural features. The differences in idiotypic specificities and in the amino acid sequences suggest that VH19.1.2 and VH9.14.7 may differ in the conformation of CDR1 and CDR2. Combining with V kappa-Ox1 gene to generate groove-type combining sites to the single site-filling epitope of alpha(1----6)dextran, the two VH chains may require certain conformations of CDR3. Whether such conformational requirements influence the choice of J minigenes, the selection of the length of VH CDR3 and the sequences at junctions, are discussed.     

Complete amino acid sequences of the heavy and light chain variable regions from two A/J mouse antigen nonbinding monoclonal antibodies bearing the predominant p-azophenyl arsonate idiotype

The immune response to p-azophenyl arsonate (Ars) in A/J mice is dominated by a cross-reactive idiotype (CRI or IdCR). IdCR+ hybridoma proteins 1F6 and 3D10 produced in a single mouse by immunization with a monoclonal anti-IdCR antibody did not bind Ars [Wysocki, L., & Sato, V. (1981) Eur. J. Immunol. 11, 832-839]. The preservation of idiotype coupled with lack of antigen binding in the same molecules provoked an examination of their primary structures in order to localize sites involved in binding to antigen and to anti-idiotypes. The VH sequence of antibody 3D10 was determined by Edman degradation of intact chains and fragments generated by CNBr, hydroxylamine, and o-iodosobenzoic acid cleavage, by trypsin and V8 protease digestion, and by sequence analysis of mRNA. The 1F6 VH sequence was reported previously [Smith, J. A., & Margolies, M. N. (1984) Biochemistry 23, 4726-4732]. The VL sequences of 1F6 and 3D10 were determined by Edman degradation of intact chains and peptides generated by cleavage with o-iodosobenzoic acid and digestion with trypsin and chymotrypsin. Both 1F6 and 3D10 are encoded by the same VH, VK, D, and JK gene segments as are IdCR+ Ars-binding antibodies. However, 1F6 and 3D10 employ the JH4 gene segment rather than JH2. Antibodies 1F6 and 3D10 share several somatic mutations, suggesting a common clonal origin, but manifest individual mutations as well. By comparison with Ars-binding IdCR+ molecules, the substitutions in 1F6 and 3D10 likely responsible for the lack of Ars binding are localized to the heavy chain D-JH junction and/or to a substitution in light chain CDR 3.     

Antibody V(h) repertoire differences between resolving and chronically evolving hepatitis C virus infections

Despite the production of neutralizing antibodies to hepatitis C virus (HCV), many patients fail to clear the virus and instead develop chronic infection and long-term complications. To understand how HCV infection perturbs the antibody repertoire and to identify molecular features of antibody genes associated with either viral clearance or chronic infection, we sequenced the V(D)J region of naïve and memory B cells of 6 persons who spontaneously resolved an HCV infection (SR), 9 patients with a newly diagnosed chronically evolving infection (CE), and 7 healthy donors. In both naïve and memory B cells, the frequency of use of particular antibody gene subfamilies and segments varied among the three clinical groups, especially between SR and CE. Compared to CE, SR antibody genes used fewer VH, D and JH gene segments in naïve B cells and fewer VH segments in memory B cells. SR and CE groups significantly differed in the frequency of use of 7 gene segments in naïve B cell clones and 3 gene segments in memory clones. The nucleotide mutation rates were similar among groups, but the pattern of replacement and silent mutations in memory B cell clones indicated greater antigen selection in SR than CE. Greater clonal evolution of SR than CE memory B cells was revealed by analysis of phylogenetic trees and CDR3 lengths. Pauciclonality of the peripheral memory B cell population is a distinguishing feature of persons who spontaneously resolved an HCV infection. This finding, previously considered characteristic only of patients with HCV-associated lymphoproliferative disorders, suggests that the B cell clones potentially involved in clearance of the virus may also be those susceptible to abnormal expansion.     

Antibody engineering for the analysis of affinity maturation of an anti-hapten response.

The influence of structural variation, previously observed in a panel of V186.2 VH/V lambda 1-expressing anti-NP antibodies from the secondary response, on the affinity of these antibodies was examined by site-specific mutagenesis and recombinant antibody construction. A tryptophan----leucine exchange at position 33 in the VH segment of all but one of the high-affinity antibodies is the most frequently observed somatic mutation and by itself leads to a 10-fold higher affinity; all other somatic exchanges are irrelevant for affinity selection. In the single case of a high-affinity antibody without this common exchange, high affinity is mediated by a combination of mutations (including a one-codon deletion) in VH and the particular D-JH rearrangement carried by this antibody. The data indicate that the pattern of somatic diversification through hypermutation is shaped by affinity selection, but that only a single point mutation is available in the VH and the VL gene of lambda 1 chain-bearing anti-NP antibodies which by itself leads to an increase of hapten-binding affinity. Based on the analysis of two secondary response antibodies from which somatic mutations in VH and VL have been eliminated, it is also concluded that the recruitment of B cell clones into the pathway of hypermutation involves a mechanism which is not based upon affinity differences towards the antigen.