The development of memory B cells takes place in germinal centers (GC) of lymphoid follicles where antigen-driven lymphocytes undergo somatic hypermutation and affinity selection, presumably under the influence of helper T cells. However, the mechanisms that drive this complex response are not well understood. We explored the relationship between GC formation and the onset of hypermutation in response to the hapten phosphorylcholine (PC) coupled to antigenic proteins in mice bearing different frequencies of CD4+ T cells. PC-reactive GC were identified by staining frozen splenic sections with peanut agglutinin (PNA) and with monoclonal Abs against AB1-2, a dominant idiotope of T15+ anti-PC antibody. The nucleotide sequences of rearranged T15 VH1 genes were determined from polymerase chain reaction amplifications of genomic DNA from microdissected GC B cells. T15+ GC became fully developed by day 6-7 after primary immunization of euthymic mice with either PC-keyhole limpet hemocyanin (KLH) or PC-chicken gamma globulin (CGG). Yet the VH1 gene segments recovered from the primary GC as late as day 10-14 had low numbers of mutations, in contrast to responses to the haptens nitrophenyl or oxazolone that sustain high levels of hypermutation after GC formation. PC-reactive B cells proliferate in histologically typical GC for considerable periods with no or little somatic hypermutation; the signals for GC formation are independent of those for the activation of hypermutation. We then examined GC 7 d after secondary immunization with PC-KLH in euthymic mice, in nu/nu mice reconstituted with limited numbers of normal CD4+ cells before priming (CD4(+)-nu/nu) and in nu/nu mice. All of these animals develop T15+ GC after antigen priming, however, the patterns of V gene mutations in the secondary GC reflected the levels of CD4+ cells present during the primary response. VDJ sequences from secondary GC of euthymic mice were heavily mutated, but most of these mutations were shared among all related (identical VDJ joints) sequences suggesting the proliferation of mutated, memory B cells, with little de novo somatic hypermutation. In contrast, the patterns of V gene diversity in secondary GC from CD4(+)-nu/nu mice suggested that there was ongoing mutation and clonal diversification during the first week after rechallenge. The secondary GC from T cell-deficient, nu/nu mice showed little evidence for mutational and/or recombinational diversity of T15+ B cells. We conclude that the participation of CD4+ helper cells is required for full activation of the mutator in GC and takes place in a dose-dependent fashion.