Proc. Natl. Acad. Sci. USA Vol. 83, pp. 1847-1851, March 1986 Immunology

Somatic evolution of variable region structures during an immune response (somatic mutation/antibody affinity/combinational diversity/clonal selection/network regulation) LAWRENCE WYSOCKI, TIM MANSER, AND MALCOLM L. GEFTER Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139 Communicated by Herman N. Eisen, October 17, 1985

ABSTRACT Immunization of strain A mice with p- the hapten, p-azophenylarsonate (Ars). An average of50% of azophenylarsonate-conijugated protein stimulates B cells that the anti-Ars antibodies obtained from hyperimmunized strain synthesize anti-p-azophenylarsonate antibodies. A large frac- A mice are serologically related by cross-reaction with a V tion of these cells produce antibodies with variable (V) regions region-specific antiserum that defines the predominant strain encoded by a single heavy chain V gene segment together with A cross-reactive idiotype, denoted as CRI-A or IdCR (13-15). multiple combinations of diversity, heavy chain joining, light Limitations on the serologically defined idiotype as a phe- chain variable, and light chain joining gene segments. Early in notypic marker for clonally related antibodies, however, the immune response, these V regions are not somatically have prompted us to analyze V regions by determining their mutated. One of these V regions is initially expressed by only primary structures and the V gene segment combinations a minority of the responding B cells but binds p-azophenylar- encoding them. This methodology allows us to monitor the sonate with the highest affinity. After a secondary immuniza- selection of related of B cells and to tion, B cells synthesizing mutated derivatives of this single V clonally populations region dominate the response and bind p-azophenylarsonate correlate selection of these clones with the structures and with even higher afflinity than does the unmutated V region. affinities of their receptor antibody V regions. These results suggest that antigen directs both the expression of We sampled B cells participating in the anti-Ars immune the immune repertoire and the amplification of V region response by generating hybridomas during primary and diversity by a sequential process of clonal selection of B cells secondary immune responses and characterized the V region expressing receptor antibodies encoded by unmutated V genes, structures of the antibodies they secrete with respect to the induction of mutation in the V genes expressed by the selected V gene segment combinations encoding them, somatic mu- cells, and reselection of B cells expressing antibodies with tation, and affinity for Ars. Our results demonstrate a strong mutated V regions of higher affinity. correlation between clonal "success"-i.e., a B-cell clone whose Ig product constitutes a significant fraction of serum The antigen specificity of an Ig is determined by the amino antibodies-and affinity for Ars. Clonal success appears to be acid sequence ofits heavy (H) and light (L) chain variable (V) determined by both information encoded in the germ line and region domains. Antibody diversity is generated, in part, by information stochastically generated by somatic mutation. the combinational assembly of five V gene segments during the construction of complete V genes (reviewed in ref. 1). A MATERIALS AND METHODS complete H chain V domain results from the joining of VH, diversity (D), and H chain joining (JH) gene segments. Antigens. Keyhole limpet hemocyanin (KLH) was pur- Similarly, the L chain V domain is fashioned by the apposi- chased from Calbiochem, and a peptide fragment containing tion of VL and JL gene segments. Additional diversity is the amino-terminal 102 amino acids of the phage X repressor generated at the junctions between V gene segments during protein was provided by Robert T. Sauer. Both ofthese were assembly (2-4) and by an unknown mechanism resulting in conjugated with the diazonium salt of Ars as reported (16). nucleotide replacements (termed "somatic mutation") (5-8). Fusions. All fusions were performed between splenic cells The latter mechanism amplifies the potential V region diver- and the Sp2/0-Ag-14 cell line (17) as reported (18) by a sity to an almost limitless extent. modification of an earlier procedure (19). A single B cell synthesizes antibodies with homogeneous V Hybridoma Cell Lines. Primary immune response hybrid- domains and expresses some of these as cell-surface recep- omas were obtained by fusion of A/J spleen cells (on the day tors. Binding of antigen to these receptor antibodies consti- indicated in Table 1) after a single intraperitoneal injection of tutes part of a signal that induces proliferation, resulting in a 300 pug of antigen in complete Freund's adjuvant. For the population that synthesizes antibodies encoded by a uniform production of secondary immune response hybridomas, A/J combination of V gene segments (9-12). While a general mice were immunized intraperitoneally with 250 gg of Ars- scheme of somatically generated antibody diversity has been KLH in complete Freund's adjuvant, were administered a elucidated, we lack an understanding of how this diversity is booster 25 days later with Ars-KLH (200 ug) in saline, and utilized during the initial process of clonal selection and the were sacrificed 3 days later for splenic fusion. subsequent acquisition of functional immunity. Initial at- Serological Assays. IdCR-bearing and Ars-binding antibod- tempts to identify clonally related V regions during an ies were detected in standard RIAs (20). immune response were dependent exclusively on serological methods (idiotype analysis). Abbreviations: Ars, p-azophenylarsonate; KLH, Keyhole limpet The idiotype investigated here is reproducibly elicited in hemocyanin; L, light; H, heavy; V, variable; VH, H chain V segment; strain A mice by immunization with proteins conjugated with VK, kappa L chain V segment; D, diversity segment; JH, H chain joining segment; CK and JK, K L chain constant and J segments; IdcR, strain A major cross-reactive idiotype; VHIdCR, the VH that encodes The publication costs of this article were defrayed in part by page charge Id' antibodies; VKIdCR, the VK that encodes the dominant class ofIdcR payment. This article must therefore be hereby marked "advertisement" antibodies; DI, the D segment of the dominant class of IdcR antibod- in accordance with 18 U.S.C. §1734 solely to indicate this fact. ies; kb, kilobase pair(s). 1847 Downloaded by guest on September 30, 2021 1848 Immunology: Wysocki et al. Proc. Natl. Acad. Sci. USA 83 (1986) Affinity Determinations. Fluorescence quenching measure- ed that seven ofthe nine hybridoma antibodies are apparently ments were performed at 220C with Ars-coupled N-acetyl-L- encoded by a single combination of V gene segments: tyrosine as reported (21). VHIdCR JH2, VKIdcR, JK1 and a D segment (DCR) whose Nucleic Acid Probes. A probe from the VH gene segment product has the deduced amino acid sequence Ser-Xaa-Tyr- that encodes IdfR antibodies, which we call VHIdc1, was Tyr-Gly-Gly-Ser-Tyr-Xaa, where Xaa denotes positions of isolated as a 0.133-kilobase-pair (kb) Ava II-Rsa I fragment variability that occur at junctions of V segment assembly encoding amino acids 15 through 59 (8). A JH (0.7 kb) probe (Fig. 1). The differences between these seven hybridoma was isolated as the nearest Xba I-EcoRI fragment located 3' proteins and the remaining two were confined exclusively to ofthe JH4 gene segment. A 0.9-kb K L chain VK probe (kindly the D region. provided by M. Shapiro and M. Weigert) covering the 5' 276 The V gene segment combination encoding the seven nucleotides of the VKio gene segment and 0.63 kb of 5' homologous V regions characterized here agrees with previ- flanking sequence was used to identify the VK segment that ous amino acid sequence data obtained from the analyses of encodes the dominant class of IdCR antibodies, VKIdCR. A J, hybridoma antibodies isolated during secondary immune probe (2.6 kb) covering the entire cluster ofJK gene segments responses to Ars on the basis of serological criteria alone: of was provided by T. Imanishi-Kari. A probe for the K L chain 16 IdCR antibodies examined, all were found to use the constant (C) region gene (Coj was provided by A. Bothwell. VHIdCR and JH2 gene segments (8, 25, 26, 28, 29); of 9 For use in hybridization analyses, all probes were labeled sequenced IdCR D regions, 8 correspond to the DCR sequence with 32P by nick-translation to a specific activity of 1-2 x 101 (26, 30); 13 IdCR antibodies possess extremely homologous cpm4&g. amino-terminal VK amino acid sequences, and 8 completely Hybridization Analyses. For the purpose of identifying sequenced chains have identical JK amino acid sequences (28, hybridomas transcribing the vHIdC gene segment, whole- 30-32), consistent with the use of either the Jhi or JK2 gene cell lysates from hybridomas grown in replicate 96-well segments. By these combined results, we estimate that microtiter trays were immobilized on nitrocellulose and between 701% and 95% of the antibodies that are synthesized hybridized with the VHIdCR probe as described (18). during the secondary immune response to Ars and that are VH and JH gene segment usage was determined by hybrid- partially encoded by the VHIdCR gene segment are also ization of VHIdCR and JH probes to EcoRI-digested hybrid- encoded by the DCR, JH2, VKIdCR, and JK1 gene segments. oma DNAs as reported (22) by a modification of an earlier Because this combination of V gene segments apparently procedure (8). encodes the serum antibodies that have been identified as the The expressed VKIdCRJK, gene segment combination was major IdCR, we use the term "canonical" to refer to it and the identified by hybridization of HindII- or BamHI-digested IdCR V region structures it encodes. hybridoma DNAs with VK1o and J, probes. The conditions Somatic Mutation of the Dominant V Regin. Knowledge of were identical to those used for VHIdCR analysis with the the primary structure of the VjIdcR germ-line gene segment exception that the hybridization and washing temperatures (8) permitted us to identify somatic changes in expressed were reduced from 520C to 440C. VHIdCR gene segments; 1200 base pairs of VHIdCR sequence Total cellular RNA from hybridomas was extracted by the from nine hybridomas isolated during the anti-Ars secondary guanidine isothiocyanate method, glyoxylated, electropho- immune response contained 37 nucleotide replacements (Fig. resed in agarose gels, transferred to nitrocellulose, and 2), consistent with our earlier conclusion based on protein hybridized with the CK and VK10 probes by established sequencing analyses that IdcR antibodies isolated during the procedures (23). Hybridization and washing conditions were identical to those used forgenomic VK gene segment analysis. RNA Sequence Analysis. Sequences of hybridoma Ig VHIdCR D JH I mRNAs were done by the method of Kaartinen et al. (24). 95 96 97 98 99 1OOMOoMOb1OOc IdCR Ser * Tyr Tyr Gly Gly Se( Tyr * CANONICAL ....AGA TCN ... TAC TAT GGT GGT AGC TAC... m... JH2 GSu Tyr RESULTS AND DISCUSSION hVH65-207 ....m6 --T GAA--T - ---TAC ---...-JH2 Domination of the Strain A Secondary Immune Response to hVH65-208 ....Am --T GNC- --N --T ---TAC JH2 Asp Tyr Ars by B Cells Expressing Antibodies Encoded by a Single hVH65-210 --G GAT --AC ... JH2 GIu Combination of V Gene Segments. We generated hybridomas ------... JH2 3 days after a secondary immunization of A/J mice with an hVH65-211 I....A --GGAA--- --N --- AGN Ars-conjugate ofKLH and selected them on the basis oftheir hVH65-212 I....AGA --G GAA-- -AN- - - - - AGN ... JH2 expression of the VHIdCR gene segment. Such a sample hVH65-213 I.... --C C N--- TGN _... JH2 His Tyr should include any hybridoma secreting either IdCR antibod- hVH65-217 --C CAT ------T TAC --- .. . JH2 ies or antibodies partially encoded by the VHId gene ....AAGAI segment but undetectable serologically by virtue of somatic Lev Met Tyr mutation or by being additionally encoded by particular D, hVH65-209 ....AGA --G GNA--T CT- ATG --G GNA -GNT --.. JH2 JH, and VL gene segments that prohibit expression of IdcR determinants. The screening technique used allows hybrid- hVH65-219 ....AGA --GAG6GA - -T GGCTAC TNC ---- J2 omas transcribing a particular V gene to be identified by homology of mRNA in whole-cell lysates with a DNA probe FIG. 1. Representative V gene segment compositions encoding (18). Of -.4000 hybridomas from three A/J mice screened VHId _Utilizing hybridoma antibodies derived from the secondary with the VHIdCR DNA probe, 75 gave positive hybridization immune response. Dashes indicate identity with the top (canonical) signals, 55 ofwhich secreted antibodies specific for Ars. The sequence; spaces are introduced to align sequences of different V gene segments encoding 9 of these monoclonal anti-Ars length. "N" denotes ambiguities in the sequence. Canonical D were regions share a 6-amino acid codon core flanked by two positions of antibodies further characterized. variability (26, 27) [96 and lOOc Kabat numbering (28)], the second An mRNA sequencing method was used to determine the of which defines the junction between the D and JH gene segments primary structure ofthe expressed D, JH and VHIdc' regions and is arbitrarily assigned here to D. The first position of variability (24). A VK probe, homologous to the VK region encoding a follows a serine codon that is not a part of VHIdCR and may result prototype IdCR V region, and a J!K probe were used in from D-D joining events. All of these antibodies are partially hybridization analyses (8, 22) to evaluate the expressed VK encoded by the canonical VJ, gene segments as determined by RNA and JK gene segments. Results of these analyses demonstrat- hybridization analyses. Downloaded by guest on September 30, 2021 Immunology: Wysocki et al. Proc. Natl. Acad. Sci. USA 83 (1986) 1849

+;;,D~~~*5 60 primary immunization. Twenty-three hybridomas secreting VHId CRGerm line TGG ATT GGA TAT ATT AAT CCT GGA AAT GGT TAT ACT AAG TAC Most Homrolo cous T GC monoclonal anti-Ars antibodies were isolated from 10 mice Secondary Response Hybridomas within 9 days after primary immunizations with Ars conju- hVH65-207 TT--AC IT gates of carrier proteins; 20 of these hybridomas were hVH65-208 -- T--GC selected by a screening method that detects antibodies hVH65-209 - -T-CC - hVH65-210 ----CA reacting with an anti-IdCR antiserum, whereas 3 ofthem were hVlH65-211 T selected by mRNA sequence homology to a VHIdCR DNA hVH65-212 C------CA- hVH65-2i3 probe. hVH65-217 The 23 hybridomas were subjected to the same analysis of hVH65-219 V gene segment expression as were the secondary immune PrImory f.Olpoisie Kybidaos... response are ; hybridomas. As expected, all ofthem expressed P6546-6PBSS6C - by the VHIdCR gene segment. Surprisingly, only 5 of the 23 primary immune response hybridoma antibodies are encoded 1P586-1 by the canonical combination of V gene segments (see Table 1). No single alternative combination encodes a large fraction of the remaining 18 antibodies. At least 10 distinct combina- * 70 AAT GAG AAG TTC AAG GGC AAG ACC ACA CTG ACT GTA GAC AAA TCC TCC AGC ACA GCC tions of V segments were identified, including a minimum of CGTc C 3 different VK gene segments, 6 different D segments (see Fig. 3), and each of the 4 JH gene segments. Thus, B cells expressing antibodies with V regions that are -G CAT- --C -G T- C ------encoded by the canonical V gene segment combination are T Table 1. Diverse combinations of V gene segments encode V region structures of antibodies produced during the primary ____- -C A- --G- _ ------~~~~~~~~~~~T- immune response to Ars Hybridoma Canonical segmentst designation* D JH (JH2) VICJK$ Overall P65D6-1 (IgG3) ND No (JH3) No§ No P65D6-2 (IgM) No No (JH4) Yes No P65D6-3 (IgM) Yes Yes Yes Yes 80 90 P65D6-4 (IgG1) ND No (JH4) Yes No TAC ATG CAG CTC AGA AGC CTG ACA TCT GAG GAC TCT GCA GTC TAT TTC TGT GCA AGA P65D6-5 (IgG2a) Yes Yes Yes Yes P65D6-6 (IgG1) Yes Yes No$ No P65D6-7 (IgG2a) Yes Yes Yes Yes --C P65H3-1 (IgM) No No (JH3) No§ No P65H3-2 (IgM) No No (JH3) No§ No P65K4-1 (IgG3) ND No (JH4) Yes No P65C6-1 (IgM) No No (JH4) Yes No P65C6-2 (IgG1) No No (JH4) Yes No P65C6-3 (IgG2b) ND No (JH3) Yes No P65J4-1 (IgG3) Yes Yes Yes Yes P65L9-1 (IgGl)'1 No No (JH1) Yes No P65L9-2 (IgGl)1 No No (JH1) Yes No P65L9-3 (IgG2a)11 ND No (JH1) No§ No P6514-1 (IgM) No Yes No§ No FIG. 2. Sequences of expressed VHIdCR genes from primary and P6514-2 (IgG3) Yes Yes Yes Yes secondary immune response hybridomas. Lines indicate identity P65B6-1 (IgM) No No (JH3) No§ No with the germ-line VHIdCR sequence. Gaps indicate ambiguity in P65A7-1 (IgM)** ND No (JH4) Yes No sequence. "Most homologous" refers to two germ-line VH gene P65G7-1 (IgGl)** No No (JH4) Yes No segments most homologous to VHIdCR, which are apparently not the P65G7-2 (IgGl)** ND No (JH4) Yes No source ofnucleotide replacements (see ref. 8). The boundaries ofthe second complementarity region are marked by arrows. ND, not determined. *AU hybridoma antibodies are encoded by the VHIdCR gene segment. secondary immune response are somatically diversified (8). Alphabetic symbols within the names denote the animals from The extent of nucleotide substitution (3%) occurring over a which the hybridomas were obtained. Numerals following alpha- short time betic symbols and preceding hyphens denote the day following (28 days since the first immunization) implies that immunization during which cell fusion was performed. V genes are apparently subjected to a high rate of somatic t'"Canonical segments" refers to the V gene segments encoding the nucleotide replacement as previously reported (33). V region structure that dominates the secondary immune response Our observation that the VL region of these antibodies is to Ars. apparently encoded by a single V,,JK gene segment combi- *The canonical VL gene is found on rearranged HindIII and BamHI nation indicates that the sequence heterogeneity in IdCR VL restriction fragments 4.3 and 9.8 kb long, respectively. domains, reported earlier (28, 30-32), is also somatically §VK gene segment is not homologous to VKIdCR. generated. IV, gene segment is homologous to VCIdCR but resides on a 2.1-kb Expression of the Dominant V Region by Only a Minority of HindIII fragment. B Cells at the 1Screened by the criterion of production of mRNA homologous to Onset of the Immune Response. We asked the VHIdCR DNA probe. The P65L9-3 (IgG2a) hybridoma is IdCR whether the one combination encoding the V region structure negative. that dominates the anamnestic response to Ars is selected **Derived from mice injected with an Ars-conjugate of the amino- from the preimmune repertoire in one step. To address this terminal 102 residues of the phage X repressor protein; the others question, we examined the immune response early after a were from Ars-KLH-immunized animals. Downloaded by guest on September 30, 2021 1850 Immunology: Wysocki et al. Proc. Natl. Acad. Sci. USA 83 (1986)

VHIdCR D JH canonical IdCR V region is based on affinity for Ars, this V 97 98 99 100 lQOolOOblOOc region should bind Ars with the highest affinity among those IdCR 94 95 96 Ser Tyr Tyr Gl Gly Ser Tyr * expressed by B cells initially participating in the response. CANONICAL..... AGA TCN... TAC TAT GGT GGT AGC TAC ... TTT JH2 His Tyr We determined the affinities for Ars-tyrosine of the 20 P65D6-3 ....AGA --C CAT- TAC -... JH2 primary anti-Ars-KLH hybridoma proteins by fluorescence Asp Tyr P65D6-6 ....AGA --G GAT- TAC -... JH2 quenching analysis (36) (Table 2). In addition, we performed Trp lie Asp Tyr Asp Alo Gly the same analysis concomitantly on the V region of a P65D6-2 ..AGA --C TGG ATT ---A- TAC GA- GCG GGC TAT GCT A-G JH4 previously isolated IdCR antibody (called 36-65), which is Arg Asn Asp Tyr Ser Vol

.... of P65C6-2 ... AGA --G AGG A---A- TA- -C- GTA TAT GCT A-G JH4 apparently encoded by the somatically unmutated version Tyr Tyr Gly Ser AGC TAT GCT A-G JH4 the canonical V gene segment combination (8). The results P65G7-1 AGA --C TAC --T --C TA- ---GT showed that none of the V regions had affinities for Ars Ser Asp Tyr Gly Gl y P65L9-2 ...AGA --G TCC- -A- TAC G-G GGC TGG TAC --C .... JH1 significantly higher than that of germ-line 36-65. Of the four lie Ser Tyr Snr three P65C6-1 ...AGA --GATC------A---NCTA- TCT TAC TAT GCTA-G JH4 V regions with affinities comparable to that of 36-65, were encoded by the canonical V gene segment combination. TGG JH3 These represented three of the five canonical V regions tested P65H3-1 ... AGA --C --- ..... Lys from one another at of V --- .... (which differ junctions segment P6514-1 . ..AGA AAG TAC JH2 may assembly and need not have identical affinities). Thus, B cells that eventually dominate the anti-Ars im- FIG. 3. Multiple D and JH gene segmenltS utilized by IdcR mune response but are only a minority of the population that hybridoma antibodies derived from the primar-y immune response. initially participates in this response express a V region See the legend to Table 1 for the origins of the hLybridoma antibodies structure that has the highest affinity for Ars. and the legend to Fig. 1 for an explanation of'the format. With the Increased Affminty of Somatically Mutated V Regions. If exception of P65H3-1 and P65I4-1, the last co onical D segments are arbitrarily assigned to th e second position of affinity for antigen directs the clonal selection process, a variability. Gaps are introduced to permit aligniment of sequences of sizeable fraction of mutated V regions expressed by B cells different lengths. participating in the secondary immune response should bind Ars with higher affinity than do their germ-line predecessors. We found that five of the six canonical antibodies from the not a majority at the onset of the immurie response. This, secondary immune response had affinities significantly high- together with the results obtained from h)ybridomas isolated er than that of 36-65 (the sixth had an affinity indistinguish- during the secondary immune response, indicates that, dur- able from 36-65). The six mutants had an average affinity 4 ing the acquisition of immunity to Ars, B1cells that express times that of 36-65 (Table 2). These results are in agreement antibodies encoded by the canonical V g,ene segment com- with studies based on hybridomas isolated by serological bination are favorably selected over others that express IdcR criteria alone (21, 34, 37). It is unlikely that the higher affinity antibodies encoded by alternative V gene segment combina- of the mutants isolated here can be attributed to an artifact of tions. the method by which the hybridomas were screened, since Absence of Somatic Mutations in V Genies Expressed by B they were chosen only on the basis of production of mRNA Cells During the Early Primary Immune Response. Sequences covering =1500 base pairs ofmRNA, specified by the VHIdCR Table 2. Association constants for Ars of hybridoma antibodies andJH gene segments expressed by 16 ofthe primary immune obtained from primary and secondary immune responses response hybridomas, revealed no somatic mutations. Con- sidering the frequency of mutations observed in the second- Hybridoma ary hybridomas, isolated only 28 days after first immuniza- antibodies Ka,*M-1 tion, we would have expected 7-10 somatic nucleotide Primary immune response replacements. This result is in agreement with that of Grif- P65D6-5t 5.6 x 105 fiths et al. (34) and with our earlier conclusion that, prior to P65D6-7,t P65D6-3t 4.5 x 1i0 immunization, expressed gene segments are apparently not P65L9-2 4.0 x 105 somatically mutated (35). P65J4-1,t P65C6-3 1.0 x 105 Thus, while IdCR V regions expressed during the early P65C6-2 8.6 x 104 primary immune response are a diverse set by virtue of the P65C6-1, P65K4-1 7.5 x 104 multiplicity of V gene segment combinations encoding them, P65D6-4 -6 x 104 they appear to lack the heterogeneity created by somatic P65D6-2, P65L9-1 -5 x 104 mutation. Because many of the primary hybridomas were P6514-1, P65I4-2t <5 x 104 isolated a week after immunization and synthesize antibodies P65L9-3, P65D6-6 <5 x 104 with C regions (IgG) indicative of previous isotype switch P65H3-1, P65H3-2 <5 x 104 recombination events, it is possible that an initial encounter P65D6-1, P65B6-1 <5 x 104 with antigen results in clonal expansion prior to the onset of Secondary immune response* somatic mutation. Perhaps such a period of B-cell growth, hVH65-212t 3.6 x 106 apparently unmarked by somatic mutation, serves the pur- hVH65-217t 2.1 x 106 pose of generating a sufficiently large pool of cells to hVH65-208t 1.6 x 106 guarantee that a random process of nucleotide replacements hVH65-207t 1.1 x 106 will produce at least one successful variant. hVH65-211t 1.0 x 106 High Affinity for Ars of the V Region That Eventually hVH65-209 9.5 x 105 Dominates the Immune Response. Because most immune hVH65-210t 4.5 x 10' responses are characterized by an increase in antibody hVH65-219 1.2 x 10' affinity as the response progresses, we were interested in the *The intrinsic association constant for Ars. The Ka determined here comparative affinities of the canonical vs. noncanonical V for 36-65 is 4.1 x 105 M-1. regions, since the former is expressed by only a minority of tThese antibodies are encoded by the canonical combination of V the responding B cells at the onset ofthe response but comes gene segments defined in the text. to be expressed by the dominating B-cell population as the tAntibodies derived from different mice than were the primary response progresses. If selection of B cells expressing the antibodies. Downloaded by guest on September 30, 2021 Immunology: Wysocki et al. Proc. Natl. Acad. Sci. USA 83 (1986) 1851 homologous to the VHIdCR DNA probe. Thus, their high 9. Burnet, F. M. (1957) Aust. J. Sci. 20, 67-69. affinity argues for a process of affinity-directed expansion of 10. Henry, C., Kimura, J. & Wofsy, L. (1972) Proc. Natl. Acad. B-cell clones that express mutationally diversified V regions. Sci. USA 69, 34-36. 11. Julius, M. H. & Herzenberg, L. (1974) J. Exp. Med. 140, The Somatic Evolution of a Humoral Immune Response. 904-920. Our results show a strong correlation between B-cell "suc- 12. Kohler, G. & Milstein, C. (1975) Nature (London) 256, cess" in this immune response and the affinities for Ars ofthe 495-497. antibodies synthesized by these cells. The V region that 13. Kuettner, M. G., Wang, A.-L. & Nisonoff, A. (1972) J. Exp. dominates this immune response is of the highest affinity Med. 135, 579-595. within a group of V regions that are elicited in unmutated 14. Milner, E. & Capra, D. (1982) J. Immunol. 129, 193-199. form at the onset of the response. B cells synthesizing this 15. Marshak-Rothstein, A., Siekevitz, M., Margolies, M. N., particular V region are selected despite the structural relat- Mudgett-Hunter, M. & Gefter, M. L. (1980) Proc. Natl. Acad. edness among members ofthis group (many are serologically Sci. USA 77, 1120-1124. 16. Nisonoff, A. (1967) in Methods in Immunology and Im- indistinguishable and share >95% sequence homology), munochemistry, eds. Williams, C. A. & Chase, M. W. (Aca- which should provide common idiotypic targets for putative demic, New York), Vol. 1, pp. 120-126. homeostatic regulation (i.e., network regulation; reviewed in 17. Shulman, M., Wilde, C. D. & Kohler, G. (1978) Nature ref. 38). In addition, secondary B cells that express somati- (London) 276, 269-270. cally mutated versions of this V region produce antibodies 18. Manser, T. & Gefter, M. L. (1984) Proc. Natl. Acad. Sci. USA that bind Ars with a higher affinity than do their unmutated 81, 2470-2474. analogues. Random somatic mutation without affinity selec- 19. Gefter, M. L., Margulies, D. H. & Scharff, M. D. (1977) tion should more frequently generate V regions with affinities Somatic Cell Genet. 3, 231-236. that are either unchanged or reduced (39). Hence, putative 20. Woodland, R. & Cantor, H. (1978) Eur. J. Immunol. 8, 600-606. simultaneous idiotype regulation resulting in the selection of 21. Rothstein, T. L. & Gefter, M. L. (1983) Mol. Immunol. 20, secondary B cells is not seen. Further evidence for affinity 161-168. selection of secondary B cells comes from a detailed exam- 22. Wysocki, L. J., Margolies, M. N., Huang, B., Nemazee, ination of the location and nature of nucleotide replacements D. A., Wechsler, D. S., Sato, V. L., Smith, J. A. & Gefter, seen in Fig. 2: repeated generation of identical amino acid M. L. (1985) J. Immunol. 134, 2740-2747. substitutions, primarily in the second complementarity- 23. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular determining region, and the paucity of silent nucleotide Cloning: A Laboratory Manual (Cold Spring Harbor Labora- changes are evident. tory, Cold Spring Harbor, NY). Immunity is generally characterized by high-affinity anti- 24. Kaartinen, M., Griffiths, G. M., Hamlyn, P. H., Markham, A. F., Kaijalinen, K., Pelkonen, J. L. T., Makela, 0. & bodies. Studies performed on whole serum antibodies by Milstein, C. (1983) J. Immunol. 130, 937-945. Eisen and Siskind (40, 41) and others (42, 43) have implicated 25. Slaughter, C. A. & Capra, D. (1983) J. Exp. Med. 158, the role of antigen as the selective agent in the affinity 1615-1634. maturation process. Our results provide a molecular descrip- 26. Gridley, T., Margolies, M. N. & Gefter, M. L. (1985) J. tion of this phenomenon. Antigen selects not only from the Immunol. 134, 1236-1244. diverse pool of unmutated V region structures expressed by 27. Kabat, E. A., Wu, T., Bilofsky, H., Reid-Miller, M. & Perry, the preimmune B-cell population but also from the additional H., eds. (1983) Sequences ofProteins ofImmunological Inter- diversity apparently generated by stochastic mutation of est (National Institutes of Health, Bethesda, MD). these V regions, resulting in a highly adaptive process in 28. Siegleman, M. & Capra, D. (1981) Proc. Natl. Acad. Sci. USA which many structural solutions to a common functional 78, 7679-7683. 29. Alkan, S. S., Ball, R. K., Chang, J. Y. & Braun, D. M. (1983) problem (i.e., achieving high affinity) are reached. From a Mol. Immunol. 20, 203-211. finite preimmune repertoire, less than or equal to the number 30. Slaughter, C., Milner, E., Estess, P., Tucker, P. W. & Capra, of B cells in the host, functional antibodies of seemingly D. (1982) Immunol. Today 3, 332-339. infinite variability are thus generated. 31. Margolies, M. N., Marshak-Rothstein, A. & Gefter, M. L. (1981) Mol. Immunol. 18, 1065-1077. We thank Shu Huang for help in sequencing analyses; Drs. 32. Ball, R. K., Chang, J. Y., Alkan, S. S. & Braun, D. G. (1983) Marjorie Shapiro, Martin Weigert, Thereza Imanishi-Kari, and Mol. Immunol. 20, 197-201. Alfred Bothwell for their contributions of DNA probes; Dr. Herman 33. McKean, D., Huppi, K., Bell, M., Staudt, L., Gerhard, W. & Eisen for the use of his spectrophotofluorimeter; and Dr. Tom Weigert, M. (1984) Proc. Natl. Acad. Sci. USA 81, 3180-3184. Rothstein for his advice on fluorescence-quenching analyses. The 34. Griffiths, G. M., Berek, C., Kaartinen, M. & Milstein, C. work was supported by a grant from the American Cancer Society (1984) Nature (London) 312, 271-275. (NP-6N to M.L.G.), grants from the National Institutes of Health 35. Manser, T., Huang, S. Y. & Gefter, M. L. (1984) Science 226, (A113357 and CA28900 to M.L.G.), and fellowships from the Amer- 1283-1288. ican Cancer Society (PF2200 to L.W. and PF2240 to T.M.). 36. Velick, S. F., Parker, C. W. & Eisen, H. N. (1960) Proc. Natl. Acad. Sci. USA 46, 1470-1482. 1. Honjo, T. (1983) Annu. Rev. Immunol. 1, 499-528. 37. Rodwell, J. D., Gearhart, P. J. & Karush, F. (1983) J. Immu- 2. Sakano, H., Huppi, K., Heinrich, G. & Tonegawa, S. (1979) Nature (London) 280, 288-294. nol. 130, 313-316. 3. Weigert, M., Perry, R. & Kelley, D. (1980) Nature (London) 38. Bona, C. & Kohler, H., eds. (1983) Ann. N. Y. Acad. Sci. 418. 283, 497-499. 39. Manser, T., Wysocki, L. J., Gridley, T., Near, R. I. & Gefter, 4. Alt, F. W. & Baltimore, D. (1982) Proc. Nad. Acad. Sci. USA M. L. (1985) Immunol. Today 57, 94-101. 79, 4118-4122. 40. Eisen, H. N. & Siskind, G. W. (1964) Biochemistry 3, 5. Crews, S., Griffin, J., Huang, H., Calame, K. & Hood, L. 996-1008. (1981) Cell 25, 59-66. 41. Eisen, H. N. (1966) in The Harvey Lectures (Academic, New 6. Kim, S., Davis, M., Sinn, E., Patten, P. & Hood, L. (1981) York), Series 60, pp. 1-33. Cell 27, 573-581. 42. Nussenweig, V. & Benacerraf, B. (1967) J. Exp. Med. 126, 7. Selsing, E. & Storb, U. (1981) Cell 25, 47-58. 727-743. 8. Siekevitz, M., Huang, S. Y. & Gefter, M. L. (1983) Eur. J. 43. Goidl, E. A., Paul, W. E., Siskind, G. W. & Benacerraf, B. Immunol. 13, 123-132. (1968) J. Immunol. 100, 371-375. Downloaded by guest on September 30, 2021