Somatic hypermutation: How many mechanisms diversify V region sequences?

lmmunoglobulin variable (V) region sequences are tailored to recognize antigen by the process of somatic hypermutation. Somatic hypermutation of V regions can occur either before or after challenge with antigen. When somatic hypermutation occurs before challenge with antigen, the result is to increase the diversity of the preimmune repertoire. When somatic hypermutation occurs in response to antigen stimulation, it is coupled with selection for antigen binding, and the result is to increase antibody affinity for a specific antigen. Insights into mechanism can sometimes come from studying related processes in a variety of organsims. Here I review some similarities between antigen-independent and antigen-driven somatic hypermutation and suggest that variations upon a single molecular mechanism might produce the distinct patterns of templated and untemplated mutation that characterize somatic hypermutation in different organisms. Antigen-Driven Somatic Hypermutation: Mutation Coupled with Selection Somatic hypermutation has been most intensively studied in mice, where hypermutation occurs after challenge with antigen and targets single base changes to the rearranged V regions. The rate of hypermutation approaches 10m3 per base pair per generation, some 105-fold higher than the mutation rate for an untargeted locus in the same cell. Hypermutation is coupled with selection for antigen binding, and a loto lOO-fold increase in affinity for specific antigen distinguishes the very good antibodies of the primary response (Kd, 1 Om7 M) from the extraordinary antibodies of the secondary response (Kd, lo-@ to 10m9 M). This dramatic increase in affinity can result from as few as two or three amino acid substitutions in a 100 residue V domain. Antigen-driven somatic hypermutation occurs in highly organized microenvironments, called germinal centers, where hypermutation is coupled with selection for antigen binding. Germinal centers develop in the follicles of the peripheral lymphoid organs following challenge with antigen (reviewed by Nossal, 1991; MacLennan, 1994). Visualized in sections of lymph node or spleen, germinal centers consist of a mantle surrounding a histologically distinct dark zone and a light zone. Small, resting B cells compose the mantle, and these cells express a large and diverse repertoire of unmutated V region sequences. B cells selected for antigen recognition populate the dark zone, where they proliferate and where somatic hypermutation occurs. Descendants of dark zone B cells migrate to the light zone, cease proliferating, and reveal their newly altered surface immunoglobulin molecules. Antigen displayed on the web of follicular dendritic cells in the light zone can then mediate affinity selection. Distinct surface markers characterize germinal center B cells at different stages of development, and this has recently permitted fractionation of germinal center B cells into distinct populations (Pascual et al., 1994) a critical step in studying both the biology and biochemistry of hypermutation. In addition to B cells, a few T cells also reside within the germinal centers. These T cells were long thought to regulate B cell hypermutation while being immune to the hypermutation process themselves. Recently, however, Kelsoe and collaborators reported that Va (but not V(3) regions of the T cell receptor genes undergo hypermutation in germinal centers (Zheng et al., 1994). T ceil receptor hypermutation is surprising and, not surprisingly, controversial (see Bach1 et al., 1995; Kelsoe et al., 1995). Although immunoglobulin gene hypermutation can be readily rationalized, T cell hypermutation seems dangerous-if these cells return to the periphery, they may exhibit new and possibly autoreactive specificities. Hotspots for Hypermutation In hypermutated murine V region sequences, mutations are not evenly distributed throughout the V region but are concentrated in the complementarity determining regions (CDRs), which encode the amino acids that make contact with antigen (see Figure 1). Clustering of hypermutation in the CDRs was noticed when the first V regions were sequenced. The ready explanation was that affinity selection must enrich for B cells carrying mutations in the regions that encode the antigen-binding site. However, it has recently been shown that targeting of somatic hypermutation to the CDRs is a property of the hypermutation mechanism itself (Betz et al., 1993a, 1993b; GonzalezFernandez and Milstein, 1993; Gonzalez-Fernandez et al., 1994; Yelamos et al., 1995). To separate the intrinsic properties of the hypermutation process from the effects of selection for antigen binding, Milstein, Neuberger, and their colleagues began by amassing a sequence database of VKOX~ light chain V regions that had undergone somatic hypermutation without affinity selection. These VK regions had escaped selection because they were carried as “passenger” trans genes that did not contribute to an antigen-specific immune response (Betz et al., 1993a, 1993b; GonzalezFernandez and Milstein, 1993). Sequences of many such hypermutated but unselected genes revealed a very strong intrinsic hotspot in CDRl of the VKOX~ region. Sub-