Genesis of the strand-biased signature in somatic hypermutation of rearranged immunoglobulin variable genes

The history and current development of the reverse transcriptase model of somatic hypermutation (RT-model) is reviewed with particular reference to the genesis of strand-biased mutation signatures in rearranged immunoglobulin variable genes (V(D)J). The recent disagreement in the field as to whether strand bias really exists or not has been critically analysed and the confusion traced to the putative presence, in some mutated V(D)J sequence collections, of polymerase chain reaction (PCR)-recombinant artefacts. Recent analysis of somatic hypermutation in xeroderma pigmentosum variant patients, by the group of PJ Gearhart and others, has established that the Y-family translesion DNA repair enzyme, DNA polymerase η (eta), is responsible for the striking A-T targeted strand-bias mutation signature seen in all mouse and human collections of somatically mutated V(D)J sequences. This evidence, together with our own recent demonstration that human DNA polymerase η is a reverse transcriptase, leads to the conclusion that the strand-biased A-T mutation signature is caused either by: (i) error-prone DNA-dependent DNA repair synthesis by pol-η of single-strand nicks preferentially in the non-transcribed strand; and/or (ii) by error-prone cDNA synthesis of the transcribed strand by pol-η using the pre-mRNA as the copying template, primed by the nicked transcribed DNA strand, followed by replacement of the original transcribed strand by cDNA. Analysis of the total mutation pattern also suggests that the major transitions observed in SHM (A→G, C→T and G→A) can be explained by known RNA editing mechanisms active on pre-mRNA which are then written into cDNA during synthesis of the transcribed strand by error-prone cellular reverse transcriptases such as pol-η.