In particular, in the transcribed strand, T residues accumulate fewer mutations than A despite the fact that they are a complementary pair (3C5). strand-independent component to IL2RG mutational targeting. Thus, there are two aspects of the hypermutation process that are sensitive to local DNA sequences, one that is usually DNA strand-dependent and the other that is not. During the maturation of the immune response, antibody genes hypermutate. This process, highly specific for the immune system, is characterized by the introduction of point mutations at a very high rate. It occurs only within a DNA segment of 1C2 Kb, encompassing the bulk of the V region but excluding the C. The B cells expressing the somatically mutated variants are then subjected to an antigen-mediated selection resulting in affinity maturation (reviewed in refs. 1 and 2). The frequency at which the four bases hypermutate suggests a strand bias. In particular, in the transcribed strand, T residues accumulate fewer mutations than A despite the fact that they are a complementary pair (3C5). This point has been used to suggest that the mutations occur on only one DNA strand and is consistent with many hypermutation models (3, 4, 6C9). However, it remains possible that the observed strand discrimination is usually caused, at least in part, by the nonrandom nature of hypermutation. The nonrandom distribution of intrinsic mutations is usually highlighted by warm as well as cold spots. There is formal proof that short sequence motifs are associated with warm spots (10, 11), but other interactions additionally have Finasteride been postulated to account for the diverse mutability of the same motif when found in different DNA segments (10, 12, 13) Thus, the nonrandom, sequence-dependent distribution of warm spots also could give rise to strand discrimination. It is not readily feasible to formally establish whether hypermutation targets only one or both DNA strands, but the problem can be approached indirectly because the rate of mutation of each base depends on its local environment. Finasteride In the case of Ig V genes, this environment is usually unlikely to be random. Indeed, analysis of codon usage in Ig V genes strongly indicates that their DNA sequences have evolved to ensure strategic localization of somatic hypermutation warm spots (14). However, by analysis of mutation in V gene flanking sequences or in transgenic non-Ig targets (11, 15), the pattern of nucleotide substitutions can be examined in sequences that are unlikely to have been subjected to evolutionary selection for nonrandom distribution of warm spots. Here, by using large databases Finasteride of such mutations, we contrast the mutation distributions observed with what would have been anticipated if either one or both DNA strands are hypermutation targets. MATERIALS AND METHODS Strategy of the Analysis. We analyzed the coding strand to establish the degree of correlation between the average mutation frequency of individual bases of triplets and of their inverted complements. Significant correlation is to be expected if both strands are hypermutation substrates. Thus, if both strands are targeted equally, the mutability of a given triplet around the coding strand should equal that of its inverted complement (e.g., 5-TAC and 5-GTA, respectively). Obviously, the reliability of our estimates of the mutation frequencies in each data set depends on the number of mutated sequences analyzed. Within each data set, these ranged from 37 to 224 (Table ?(Table1),1), which we assume are sufficient for meaningful conclusions. Pooling all data into a single database would have given undue weight Finasteride to the sets represented by the largest number of sequences. Thus, we separately calculated the mean mutation frequency for each base type in every triplet of our data sets, and only then were the values pooled. Table 1 Mutation?databases be the number of occurrences of a given triplet (= T, A, C, or G) in each wild-type sequence and the.
