Strand bias in Ig somatic hypermutation is determined by signal sequence within the variable region.

Abstract

Ig genes undergo hypermutation with a nucleotide preference of A over T for mutation on the coding strand. As only with concomitant strand bias can such nucleotide bias be observed, Ig gene hypermutation is generally accepted as a strand-specific process, for which the mechanistic basis remains unknown. It has previously been shown that different non-Ig sequences replacing the LVJ region of an Ig transgene to various extents are targeted for hypermutation with similar mutation frequencies. However, the nucleotide bias characteristic of Ig hypermutation was not found in two of the three such sequences studied. To test whether it is the DNA sequences of the non-Ig substrates that determine the pattern of nucleotide bias in hypermutation or whether the LVJ sequence may contain element(s) that confer strand bias, we have added back all the replaced LVJ sequences to one of the transgenes, L(kappa)-Vgpt*, that expresses no strand bias in hypermutation and studied the outcome. The results show that the gpt sequence in the presence of the complete LVJ sequence hypermutates differently from the same sequence in L(kappa)-Vgpt* where 84% of the LVJ was replaced. The main difference is the resumption of strand bias characteristic of Ig hypermutation. Thus, whether or not a substrate sequence manifests strand bias in hypermutation is not inherently determined by the substrate DNA sequence. This indicates the presence of special element(s) within the LVJ that confer strand bias.