The Journal of Immunology
Bcl-2 Rescues the Germinal Center Response But Does Not Alter the V Gene Somatic Hypermutation Spectrum in MSH2-Deficient Mice1 Boris Alabyev and Tim Manser2 Ab V genes in mice deficient for the postreplication mismatch repair factor MutS homolog (MSH2) have been reported to display an abnormal bias for hypermutations at G and C nucleotides and hotspots. We previously showed that the germinal center (GC) response is severely attenuated in MSH2-deficient mice. This suggested that premature death of GC B cells might preclude multiple rounds of hypermutation necessary to generate a normal spectrum of base changes. To test this hypothesis, we created MSH2-deficient mice in which Bcl-2 expression was driven in B cells from a transgene. In such mice, the elevated levels of intra-GC apoptosis and untimely GC dissolution characteristic of MSH2-deficient mice are suppressed. However, the spectrum of hypermutation is unchanged. These data indicate that the effects of MSH2 deficiency on GC B cell viability and the hypermutation process are distinct. The Journal of Immunology, 2002, 169: 3819 –3824.
D
uring the course of T cell-dependent B cell immune responses, the structure and function on Ab-variable regions expressed by responding B cells undergo extensive alteration due to somatic hypermutation of the genes encoding them (1, 2). Although hypermutation has been the subject of intense investigation, its mechanism remains poorly understood. Nonetheless, recent evidence suggests that components of the cellular machinery required for normal DNA replication and repair are involved in this process (3, 4). In particular, error-prone short patch DNA polymerases have been suggested to participate in the introduction of lesions during hypermutation (3, 4). Moreover, deficiencies in several factors involved in postreplication mismatch repair have been shown to alter the outcome of hypermutation in vivo (5, 6). The most intensely studied of these is MutS homolog (MSH2),3 the mammalian homolog of the bacterial mismatch repair factor MutS. During normal DNA replication, MSH2, in combination with other mismatch repair factors such as MSH6 and MSH3, binds to DNA mispairs or small deletion/insertion loops resulting from DNA polymerase errors or chemical modification (7). This binding results in signals that culminate in recruitment of other factors that ultimately repair the lesion (8). Analysis of the V genes expressed by germinal center (GC) and post-GC B cells in mice deficient in MSH2 has consistently revealed a perturbation in the spectrum of bases altered by the hypermutation process (9 –12). In contrast to normal mice, MSH2deficient mice often display a strong bias for mutations at G and C nucleotides, particularly in regions identified as hypermutation
Kimmel Cancer Center and Department of Microbiology and Immunology, Jefferson Medical College, Thomas Jefferson University Philadelphia, PA 19107 Received for publication May 1, 2002. Accepted for publication August 1, 2002. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1
“hotspot” motifs. Interestingly, Ig somatic hypermutation in ectotherms (13, 14), which lack GCs, and in certain B cell lines (15, 16) displays a bias for lesions at G and C positions. These data have suggested a model invoking perturbed “fixation” of hypermutations in mouse GCs in the absence of MSH2 (5, 6). According to this model, the mechanism responsible for introduction of mutations, or a component thereof, preferentially alters G and C bases. In the presence of MSH2, however, many of these lesions are repaired, resulting in the eventual accumulation of mutations at A and T residues as B cell undergo successive cycles of the hypermutation process in the GC. Due to the central role of postreplication mismatch repair in the maintenance of overall genome integrity, we reasoned that the absence of MSH2 might have generic debilitating effects on B cell responses in vivo, particularly during stages of lymphocyte differentiation characterized by high rates of proliferation. To test this idea, we analyzed the kinetics and magnitude of the Ab-forming cell and GC responses of MSH2-deficient mice (17). Both responses were severely attenuated. Whereas initiation of the GC response appeared normal, MSH2-deficient GCs were characterized by increased levels of apoptosis, and the response rapidly waned. On the basis of these observations, we suggested that the effect of MSH2 deficiency on the V gene hypermutation spectrum might be indirect, resulting from the loss of GC B cells that would have normally undergone multiple rounds of hypermutation. This idea was also suggested and supported by Weill’s and Neuberger’s groups, who observed a reduced overall frequency of V gene mutations in Peyer’s patch B cells of MSH2-deficient mice, accounted for by a reduction in the frequency of heavily mutated V genes (11, 12). To further test the idea that reduced viability of GC B cells might account for the hypermutation G ⫹ C and hotspot bias due to MSH2 deficiency, we created lines of MSH2-deficient mice in which the expression of the anti-apoptotic factor Bcl-2 is driven in B cells from a transgene.
This work was supported by National Institutes of Health Grant AI23739 (to T.M.).
2
Address correspondence and reprint requests to Dr. Timothy Manser, Kimmel Cancer Center, BLSB 708, 233 South 10th Street, Philadelphia, PA 19107. E-mail address:
[email protected] 3
Abbreviations used in this paper: MSH, MutS homolog; GC, germinal center; Tg, transgenic; PNA, peanut agglutinin; CDR, complementarity-determining region; BMSH2 mice, Msh2⫺/⫺Bcl-2-Tg⫹ mice. Copyright © 2002 by The American Association of Immunologists, Inc.
Materials and Methods Mice and immunizations The MSH2-deficient and Bcl-2-transgenic (Tg) lines used in this study have been previously described (18, 19). The presence of the MSH2 knockout allele and the Bcl-2 transgene were assayed in tail DNA of offspring of 0022-1767/02/$02.00
3820
GCs IN Bcl-2-Tg, MSH2-DEFICIENT MICE
crosses between these lines using PCR and Southern blotting, respectively, as previously described (20, 21). Mice used in the studies were 8 –12 wk old, and littermates were used in most experiments. Mice were immunized with 100 g of NP21-CGG (Biosearch Technologies, Novato, CA) precipitated on alum and injected i.p.
Immunohistology Spleens were isolated and flash frozen, and 5-m sections were prepared as previously described (17). Sections were stained with either peanut agglutinin-(PNA)-HRP (Sigma-Aldrich, St. Louis, MO) and the anti-1 mAb LS136-biotin, followed by streptavidin-alkaline phosphatase (DAKO, Glostrup, Denmark) and then visible dye HRP and alkaline phosphatase developing reagents (Vector Laboratories, Burlingame, CA); or GL7-FITC (BD PharMingen, San Diego, CA) and LS136-biotin followed by streptavidin-PE (Molecular Probes, Eugene, OR). In some experiments, parallel sections were first subjected to visible dye (ApopTag peroxidase in situ apoptosis detection kit; Intergen, Purchase, NY) or fluorescent (ApopTag Red in situ apoptosis detection kit; Intergen) TUNEL assays and then stained with either PNA-HRP or GL7-FITC.
Microdissection-PCR-nucleotide sequencing Tissue from individual PNA⫹ GCs in sections were microdissected and digested as previously described (22). Genomic DNA seminested PCR were then performed on these samples. V1 genes were amplified as previously described (10). VH186.2 genes were amplified using a VH186.2 leader region primer (5⬘-ACACAGGACCTCACCATG-3⬘) and first and second round primers that hybridize in a region 3⬘ of JH4 (5⬘-CCTGG AGAGGCCATTCTTACCTGA-3⬘) and a region between JH3 and JH4 (5⬘TCACAAGAGTCCGATAGACC-3⬘), respectively. PCR products of the appropriate sizes were purified on agarose gels, and cloned into plasmid vectors using the pGEM-T Easy Vector System I (Promega, Madison, WI). Plasmid inserts were subjected to DNA sequencing in the Kimmel Cancer Center Nucleic Acids Facility, and alignment with germline sequences was performed using the CLUSTAL W multiple sequence alignment program. In total, 16 clones from 15 GCs from the spleens of 2 wild-type mice, 27 clones from 25 GCs from the spleens of 5 Msh2⫺/⫺ mice, and 31 clones from 31 GCs from the spleens of 11 Bcl-2-Tg⫹Msh2⫺/⫺ mice were obtained and analyzed. Twenty-five clones from the latter mice were from day 12 and six from day 14 after immunization. All of the Msh2 ⫺/⫺ and wild-type GCs were from day 12. Because mice used in the study had heterogeneous genetic backgrounds (mixed A/J, C57BL/6, and 129) the region between JH2 and JH4 was PCR amplified and cloned from kidney DNA from each of the background strains and sequenced. In this way, nucleotide differences in the GC PCR clones in this region resulting from allelic polymorphisms were identified and were not scored as somatic mutations.
Results We previously showed that a line of mice expressing Bcl-2 from a hybrid human Bcl-2-Ig transgene in B cells (19) displays reduced intra-GC apoptosis and enlarged GC size relative to nontransgenic littermates (21, 23). To determine whether the effects on the GC reaction of Bcl-2 expression would be dominant to those of an MSH2 deficiency, Bcl-2-Tg and Msh2⫺/⫺ lines of mice were crossed, and their offspring were intercrossed to yield Msh2⫺/⫺Bcl-2-Tg⫹ mice and all other possible genetic combinations. Msh2⫺/⫺Bcl-2 Tg⫹ (BMSH2) mice and littermates of other genotypes were immunized with NP-CGG, and spleens were taken at day 12 and processed for histology. This time point was chosen because we previously observed that GCs became extremely infrequent in Msh2⫺/⫺ mice at later times (17). GCs were elaborated using PNA or GL7 staining, and apoptotic nuclei were visualized via the TUNEL assay. Tissue sections were also stained with anti1 to reveal Ag-specific GCs (because the primary anti-NP response is dominated by -expressing B cell clonotypes (24)). Fig. 1 shows that, as we previously reported, the Ag-specific splenic GC response in Msh2⫺/⫺ mice is substantially reduced as compared with that of wild-type littermates. This reduction is manifested in both the number and size of GCs, and Msh2⫺/⫺ mice essentially lack GCs larger than 30 cell diameters (medium and large GCs). In contrast, Bcl-2-Tg mice display enhanced GC responses, in terms of both size and number, as compared with
FIGURE 1. Expression of Bcl-2 rescues the GC response in MSH2deficient mice. Mice of the indicated genotypes (wild type (WT) ⫽ Msh2⫹/⫹, Bcl-2 Tg⫺) were immunized with NP-CGG as described in Materials and Methods and sacrificed at day 12 thereafter. Spleens sections were prepared and stained also as described in Materials and Methods and evaluated for numbers and sizes of ⫹ GCs. Six randomly selected ⫻64 fields (final magnification) from 2 sections at least 50 sections apart were scored from each spleen. At least three mice per genotype were evaluated. GC sizes were scored as follows: tiny ⫽ 5–20 PNA⫹ cell diameters; small ⫽ 20 –30 diameters; medium ⫽ 30 – 40 diameters; large ⫽ ⬎40 diameters. Error bars reflect the SD in values obtained from individual animals of the same genotype.
MSH2-sufficient mice lacking the Bcl-2 transgene (wild type). Analysis of BMSH2 spleen sections revealed an overall ⫹ GC response most similar to that of their Bcl-2-Tg⫹ littermates, a response that appeared more robust than that of wild-type mice. At day 18, when GCs are extremely rare in Msh2⫺/⫺ mice (17), a more limited comparison of GC size and number between wildtype and BMSH2 littermates revealed ongoing GC responses that were similar in magnitude (data not shown). TUNEL analysis of spleen sections showed that, as expected from previous studies (17), Msh2⫺/⫺ mice had significantly increased numbers of intra-GC apoptotic nuclei as compared with Bcl-2-Tg mice (Fig. 2). In contrast, BMSH2 splenic GCs displayed levels of apoptotic nuclei similar to their Msh2⫹/⫹, Bcl-2-Tg⫹ littermates. Taken together with the analysis of GC size and frequency, these data indicate that expression of Bcl-2 suppresses the elevated apoptotic response characteristic of Msh2⫺/⫺ GC B cells, resulting in quantitative restoration of the GC response. A microdissection-PCR approach was used to analyze the frequency and chemical nature of V gene somatic hypermutation at stages of the GC response defined as intermediate in normal mice, given that GCs become extremely small and rare in MSH2-deficient mice at later times. Tissue from individual ⫹ GCs was subjected to PCR using primers specific for V1, or for the VH186.2 J558 VH subfamily and flanking sequence. The VH186.2 gene is used predominantly by anti-NP clonotypes in C57BL/6 mice (24). Flanking sequence was studied to minimize the effects of antigenic selection on the mutations sampled. PCR products of the appropriate size were cloned and subjected to DNA sequencing, and the resulting sequences were compared with the germline sequences of the V1 gene, the VH186.2 gene, and JH regions. In several clones, the VH region coding sequence had a high degree of homology to previously reported analog genes expressed in the C57BL/6 antiNP response (22). Due to the uncertainty of the germline origins of the VH gene segments encoding these genes, they were excluded from the mutational analysis.
The Journal of Immunology
FIGURE 2. Expression of Bcl-2 suppresses the elevated levels of intra-GC apoptosis characteristic of MSH2-deficient mice. Mice were immunized, and spleens were processed for histology as described in the legend to Fig. 1 and in Materials and Methods. Spleen sections were additionally subjected to the immunofluorescence TUNEL assay. Due the absence of large GCs in Msh2 ⫺/⫺ mice and the restricted size of tiny GCs in all mice, the number of apoptotic nuclei could be comparatively evaluated only in small and medium-sized GCs. The scatter plots illustrate the number of apoptotic nuclei in individual GCs, and the mean values obtained for each genotype are indicated. A Student t test was used to evaluate the significance of the differences observed. Those between the values obtained for BMSH2 and Msh2 ⫺/⫺, and wild-type and Msh2 ⫺/⫺ mice were highly significant in both cases (0.013– 0.00005), whereas the differences between values obtained for BMSH2 and wild-type mice were not (0.7– 0.8). Similar results were obtained using the visible dye staining TUNEL assay.
The overall frequency of mutations in the VH186.2 genes in Msh2⫺/⫺, BMSH2 and wild-type GCs were 1.8, 1.4, and 1.2%, respectively. The frequency in DNA flanking the 3⬘ side of VH genes was 0.4% in Msh2⫺/⫺ and 0.7% in BMSH2 and wild-type
FIGURE 3. V region genes containing a high frequency of somatic mutation are more prevalent in the GCs of BMSH2 mice than in the GCs of MSH2-deficient mice. VH gene coding and flanking regions were PCR amplified, cloned from individual GCs, and sequenced, and sequences were scored for somatic mutations as described in Materials and Methods. The frequencies of clones containing the indicated numbers of mutations are illustrated by pie charts. The number of clones analyzed is indicated in the middle of each chart. Mutations were scored only in clones from different GCs or in clones from the same GC that contained different V-D-J region sequences to maximize the chance that each mutation represented an independent event.
3821 GCs. Although these bulk values indicate that the frequencies of somatic mutation are similar in V genes isolated from MSH2⫺/⫺ and BMSH2 GCs, Fig. 3 illustrates that the frequency of individual PCR clones that displayed large numbers of mutations was higher in BMSH2 GCs than in MSH2⫺/⫺ GCs, particularly in flanking regions. The level of mutation in V clones, determined only for BMSH2 GCs, was substantially lower than in VH coding and flanking sequence (0.3%). Fig. 4 illustrates that in the VH186.2-coding and 3⬘-flanking regions of PCR clones obtained from both Msh2⫺/⫺ and BMSH2 GCs, somatic mutations at G and C nucleotides are somewhat overrepresented as compared with the base composition of these regions (53 and 48% G ⫹ C, respectively), and the frequency of mutations at these bases in wild-type GCs. The levels of mutation observed in V clones obtained from BMSH2 GCs displayed this same bias for changes at G and C positions (⬎70%; data not shown). Fig. 5 shows that several of the mutations observed in VH186.2 coding region clones obtained from both Msh2 ⫺/⫺ and BMSH2 GCs were observed multiple times in the same codons in clones representing independent mutational events. Many of these codons overlap or contain RGYW or WA consensus hotspot motifs (where R ⫽ purine, Y ⫽ pyrimidine, and W ⫽ A or T) or closely related motifs (25, 26). However, no striking differences in the frequency of base changes at these positions were observed between the clones obtained from Msh2⫺/⫺ and BMSH2 GCs. Analogous results were obtained when mutations in 3⬘-flanking regions were compared (data not shown). Fig. 5 also illustrates that among somatic mutations in the VH186.2 genes isolated from BMSH2 and Msh2⫺/⫺ GCs, a position 33 Trp3 Leu mutation is frequently observed. This mutation has been previously shown to increase the affinity of VH186.2/ V1-encoded Abs for NP 10-fold (27). In clones obtained from BMSH2 GCs, the ratio of mutations causing amino acid replacements to those that do not is high in the complementarity-determining region (CDR) 1 and 2 subregions (⬃5.9:1) and low in the
3822
FIGURE 4. Somatic mutations that alter G and C bases are predominant in both the coding and flanking regions of V genes in the GCs of MSH2deficient and BMSH2 mice. Mice were immunized, spleens were processed for histology, GCs were microdissected, and VH186.2 coding and flanking regions were PCR amplified, cloned, and sequenced as described in Materials and Methods. Mutations were scored, and the percentage occurring at the indicated type of base is illustrated in pie charts. The total number of mutations sampled in each case is indicated in the center of each chart. The germline base composition of the VH186.2 coding region is 53% G ⫹ C, and this composition in the 3⬘-flanking region analyzed is 48% G ⫹ C. WT, wild type.
framework regions (2.2:1). In addition, the majority of the VH186.2 genes isolated from both BMSH2 and MSH2⫺/⫺ GCs contain CDR3 regions 9 –10 codons long (data not shown). Selection of CDR3 regions of this length is a previously recognized characteristic of the anti-NP response in wild-type C57BL/6 mice (22). Finally, we observed no mutations in the V1 or VH186.2 gene isolated from BMSH2 GCs that would directly result in a termination codon and only two that would cause shifts in translational reading frame. Collectively, these data suggest that affinitybased positive selection of hypermutating B cells is operative in BMSH2 GCs and apparently in Msh2⫺/⫺ GCs as well.
Discussion Our data demonstrate that driving expression of Bcl-2 in B cells from a transgene rescues the GC response in MSH2-deficient mice, at least up to and including its intermediate stages. Although further studies will be required to determine the mechanism of action of Bcl-2 in this regard, it seems reasonable to speculate that it blocks or delays the death of GC B cells that have undergone genome-wide lesions that would normally lead to their elimination via apoptotic pathways. Such lesions would be expected to accumulate rapidly in GC B cells in the absence of an intact mismatch repair system, given that the GC B cell cycle time is very short
GCs IN Bcl-2-Tg, MSH2-DEFICIENT MICE (28). Consistent with previous conclusions that an MSH2 deficiency precludes multiple rounds of GC B cell proliferation and mutation necessary to allow accumulation of high frequencies of mutation per V gene (11, 12), we isolated heavily mutated V gene clones from BMSH2 but not Msh2⫺/⫺ GCs. Finally, our V gene sequencing analyses also provided evidence for efficient Ag affinity-based positive selection of BMSH2 GC B cells. We previously suggested that the G ⫹ C bias observed among V gene somatic mutations in the GC and post-GC B cells of MSH2-deficient mice was due to premature termination of the GC reaction, skewing the sampling of GC and post-GC B cells to those that had undergone only the initial stages of hypermutation, perhaps mainly at hotspots (17). Contrary to the predictions of this model, however, the somatic mutations found in the V genes of BMSH2 GCs displayed a bias for G and C target nucleotides similar to that found in the V genes of Msh2⫺/⫺ GCs. In addition, we did not detect any striking differences in the frequency of mutations at V gene hotspots in Msh2⫺/⫺ and BMSH2 GCs. Taken together, these data indicate that the effects of an MSH2 deficiency on the physiology of the GC reaction and the hypermutation process are distinct. Supporting the generality of this idea for mismatch repair deficiencies, Kim et al. (29) previously reported that in Mlh1- and Pms2-deficient mice, similar proportions of GC B cells were present in spleen and Peyer’s patches as in control mice, but V gene somatic mutations were biased for those having taken place at G and C nucleotides. Apparently, even when the GC response is quantitatively diminished and attenuated, sufficient numbers of Msh2⫺/⫺ GC B cells survive to allow fairly accurate sampling of the V gene products of hypermutation at steady state via the individual GC microdissection-PCR approach. As such, our studies support previous suggestions that the outcome of the hypermutation process represents a balance between the introduction of base changes and their repair by processes such as those initiated by MSH2 (5, 6, 29). Nonetheless, questions remain regarding the level of impact of MSH2 and mismatch repair pathways in general on the outcome of hypermutation. In this regard, during the course of our analysis, we evaluated the frequency of silent (not resulting in amino acid replacements possibly subject to selection) mutations at G and C positions present in the VH186.2 gene coding region reported in previous studies of the anti-NP response of MSH2-sufficient mice. Even among studies that catalogued large numbers of mutations and V clones, these values varied from ⬃50 to 70%, despite a G ⫹ C composition in this region of 53%. We obtained values of 65 and 74% in the coding region VH186.2 genes recovered from Msh2⫺/⫺ and BMSH2 GCs, respectively. Restricting the analysis to silent mutations in the coding region yielded a value of 84% for BMSH2 GCs, and 60% of silent mutations observed in the Msh2⫺/⫺ GCs were at G or C positions. In the 3⬘-flanking regions of VH186.2 and related VH genes, where mutations should not have been influenced by antigenic selection, the values were 64 and 69%. In total, these data suggest that the influence of an MSH2 deficiency on the mutation spectrum in the VH186.2 genes expressed by GC B cells at the intermediate stages of the response to NP is rather subtle. Curiously, a previous analysis of hypermutation of VH186.2 and related VH genes focusing on day 10 of the anti-NP response in mice deficient in the postreplication mismatch repair factor MSH6 revealed that all mutations observed were at G or C positions (30). This extreme skewing agrees with the original data of Phung et al. (9) and colleagues on MSH2-deficient mice, using the immune response to oxazolone as a model. It is also consistent with the fact that MSH2 functions in mismatch repair, at least in part, in heterodimeric association with MSH6. However, in both these previous studies, suspension staining of total splenic B cells followed by
The Journal of Immunology
3823
FIGURE 5. The distributions of somatic mutations in the VH186.2 genes present in the GCs of MSH2-deficient and BMSH2 mice are not noticeably different. The location and frequency of mutations in the VH186.2 coding region of GC PCR clones obtained as described in Materials and Methods are illustrated schematically. f, Mutations resulting in amino acid replacements; 䡺, silent mutations. Mutations were scored as described in the legend to Fig. 3, to maximize the sampling of independent mutational events. One of these is due to a TGG3 TTG change at codon 33, resulting in a Trp (W)3 Leu mutation that results in a 10-fold increase in affinity for NP (27). Codons in which mutations were observed in numerous clones are indicated by arrows, and the amino acids encoded by these codons are indicated. Distributions of mutations not resulting in amino acid replacements are rather evenly distributed throughout the V genes isolated from both types of GCs.
flow cytometry was used to enrich for B220⫹, PNA⫹ cells. In contrast, we used direct sampling of PNA⫹ cells in GCs defined by histological staining of spleen sections. Given the reduction in GC B cell viability due to certain mismatch repair deficiencies, it is possible that these previous studies preferentially sampled GC B cells that had not yet initiated high rate proliferation and hypermutation, due to their enhanced survival relative to other mismatch repair-deficient GC B cells during staining and sorting procedures. Such an interpretation would be in keeping with the previous speculations of Rada et al. (11) that the initial stages of hypermutation result in predominant alteration of G and C bases, particularly those present in mutational hotspots. Finally, data from numerous laboratories indicate wide ranging, sometimes conflicting, and often less than striking effects of individual deficiencies in a variety of error-prone DNA polymerases and other DNA repair factors on the outcome of hypermutation (4, 5, 29, 31). Taken together with our results, these data indicate that the direct and indirect actions of numerous such factors culminate to produce the frequency and spectrum of base changes characteristic of hypermutation. The discovery that a deficiency in the putative RNA-editing factor termed activation-induced deaminase nearly ablates the hypermutation of Ab V genes (32) indicates that some factors are clearly more important than others in determining the outcome of hypermutation. However, because the activationinduced deaminase factor is also required for efficient Ab heavy chain class switching (32) and diversification of V genes via gene conversion in chicken B cells (33), it seems likely that it is involved in the generation of substrates or activities necessary for hypermutation, rather than the introduction of mutations per se.
Acknowledgments We thank Drs. Stanley Korsmeyer and Richard Fishel for the original strains of Bcl-2-Tg and MSH2-deficient mice, respectively. We also thank
Kate Dugan, Scot Fenn, and William Monsell for technical assistance and all other members of the Manser laboratory for their indirect contributions to this work.
References 1. Manser, T., L. J. Wysocki, M. N. Margolies, and M. L. Gefter. 1987. Evolution of antibody variable region structure during the immune response. Immunol. Rev. 96:141. 2. Berek, C., and C. Milstein. 1987. Mutation drift and repertoire shift in the maturation of the immune response. Immunol. Rev. 96:23. 3. Poltoratsky, V., M. F. Goodman, and M. D. Scharff. 2000. Error-prone candidates vie for somatic mutation. J. Exp. Med. 192:F27. 4. Storb, U. 2001. DNA polymerases in immunity: profiting from errors. Nat. Immunol. 2:484. 5. Kim, N., and U. Storb. 1998. The role of DNA repair in somatic hypermutation of immunoglobulin genes. J. Exp. Med. 187:1729. 6. Wiesendanger, M., M. D. Scharff, and W. Edelman. 1998. Somatic hypermutation, transcription and DNA mismatch repair. Cell 94:415. 7. Kolodner, R. 1996. Biochemistry and genetics of eukaryotic mismatch repair. Genes Dev. 10:1433. 8. Fishel, R. 1998. Mismatch repair, molecular switches and signal transduction. Genes Dev. 12:2096. 9. Phung, Q. H., D. B. Winter, A. Cranston, R. E. Tarone, V. A. Bohr, R. Fishel, and P. J. Gearhart. 1998. Increased hypermutation at G and C nucleotides in immunoglobulin genes from mice deficient in the MSH2 mismatch repair protein. J. Exp. Med. 187:1745. 10. Jacobs, H., Y. Fukita, G. T. L. van der Horst, J. de Boer, G. Weeda, J. Esser, N. de Wind, B. P. Engelward, L. Samson, S. Verbeek, et al. 1998. Hypermutation of immunoglobulin genes in memory B cells of DNA repair-deficient mice. J. Exp. Med. 187:1735. 11. Rada, C., M. R. Ehrenstein, M. S. Neuberger, and C. Milstein. 1998. Hot spot focusing of somatic hypermutation in MSH2-deficient mice suggests two stages of mutational targeting. Immunity 9:135. 12. Frey, S., B. Bertocci, L. Delbos, F. Quint, J.-C. Weill, and C.-A. Reynaud. 1998. Mismatch repair deficiency interferes with the accumulation of mutations in chronically stimulated B cells and not with the hypermutation process. Immunity 9:127. 13. Wilson, M., A. Marcuz, and L. D. Pasquier.1995. Somatic mutations during an immune response in Xenopus tadpoles. Dev. Immunol. 4:227. 14. Diaz, M., and M. F. Flajnik. 1998. Evolution of somatic hypermutation and gene conversion in adaptive immunity. Immunol. Rev. 162:13. 15. Green, N. S., M. M. Lin, and M. D. Scharff. 1998. Immunoglobulin hypermutation in cultured cells. Immunol. Rev. 162:67.
3824 16. Bachl, J., and M. Wabl. 1996. An immunoglobulin mutator that targets G:C base pairs. Proc. Natl. Acad. Sci. USA 93:851. 17. Vora, K. A., K. Tumas-Brundage, V. Lentz, A. Cranston, R. Fishel, and T. Manser. 1999. Severe attenuation of the B cell immune response in MSH2deficient mice. J. Exp. Med. 189:471. 18. Reitmair, A. H., R. Scmits, A. Ewel, B. Bapat, M. Redston, A. Mitri, P. Waterhouse, H.-W. Mittrucker, A. Wakeham, B. Liu, et al. 1995. MSH2 deficient mice are viable and susceptible to lymphoid tumors. Nat. Genet. 11:64. 19. McDonnell, T., N. Deane, F. M. Platt, G. Nunez, U. Jaeger, J. P. McKearn, and S. J. Korsmeyer. 1989. bcl-2 immunoglobulin transgenic mice demonstrate extended B cell survival and follicular lymphoproliferation. Cell 57:79. 20. Cranston, A., T. Bocker, A. Reitmair, J. Palazzo, T. Wilson, T. Mak, and R. Fishel. 1997. Female embryonic lethality in mice nullizygous for both Msh2 and p53. Nat. Genet. 17:114. 21. Hande, S., E. Notidis, and T. Manser. 1998. Bcl-2 obstructs negative selection of autoreactive, hypermutated antibody V regions during memory B cell development. Immunity 8:189. 22. Jacob, J., J. Przylepa, C. Miller, and G. Kelsoe. 1993. In situ studies of the primary immune response to (4-hydroxy-3-nitrophenyl) acetyl. III. The kinetics of V-region mutation and selection in germinal center B-cells. J. Exp. Med. 178: 1293. 23. Notidis, E., S. Hande, and T. Manser. 2001. Enforced expression of Bcl-2 selectively perturbs negative selection of dual reactive antibodies. Dev. Immunol. 8:223. 24. Bothwell, A. L., M. Paskind, M. Reth, T. Imanishi-Kari, K. Rajewsky, and D. Baltimore. 1981. Heavy chain variable region contribution to the NPb family of antibodies: somatic mutation evident in a ␥2a variable region. Cell 24:625.
GCs IN Bcl-2-Tg, MSH2-DEFICIENT MICE 25. Rogozin, I. B., and N. A. Kolchanov. 1992. Somatic hypermutagenesis in immunoglobulin genes. II. Influence of neighboring base sequence on mutagenesis. Biochim. Biophys. Acta 1171:11. 26. Rogozin, I. B., Y. I. Pavlov, K. Bebenek, T. Matsuda, and T. A. Kunkel. 2001. Somatic mutation hotspots correlate with DNA polymerase ⑀ error spectrum. Nat. Immunol. 2:530. 27. Allen, D., T. Simon, F. Sablitsky, K. Rajewsky, and A. Cumano. 1988. Antibody engineering for the analysis of affinity maturation of an anti-hapten response. EMBO J. 7:1995. 28. Shang, J., I. C. M. MacLennan, J. J. Liu, and P. J. L. Lane. 1988. Is rapid proliferation in B centroblasts linked to somatic mutation in memory B cell clones? Immunol. Lett. 18:297. 29. Kim, N., G. Bozek, J. C. Lo, and U. Storb. 1999. Different mismatch repair deficiencies all have the same effects on somatic hypermutation: intact primary mechanism accompanied by secondary modifications. J. Exp. Med. 190:21. 30. Wiesendanger, M., B. Kneitz, W. Edelmann, and M. D. Scharff. 2000. Somatic hypermutation in MutS homologue (MSH)3-, MSH6-, and MSH3/MSH6-deficient mice reveals a role for the MSH2-MSH6 heterodimer in modulating the base substitution pattern. J. Exp. Med. 191:579. 31. Reynaud, C. A., B. Bertocci, S. Frey, F. Delbos, L. Quint, and J. C. Weill. 1999. Mismatch repair and immunoglobulin gene hypermutation: did we learn something? Immunol. Today 20:522. 32. Masamichi, M., K. Kinoshita, S. Fagarasan, S. Yamada, Y. Shinkai, and T. Honjo. 2000. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102:553. 33. Arakawa, H., J. Hauschild, and J. M. Buerstedde. 2002. Requirement of the activation-induced deaminase (AID) gene for immunoglobulin gene conversion. Science 295:1301.
