Integration of a polyomavirus recombinant containing highly repetitive sequence : analysis of the junctions and genome rearrangements
Wallenburg, John C.
Studies on the integration of exogenous DNA into the genomes of mammalian cells have established that three different recombination mechanisms are involved; sites-pecific, illegitimate and homologous recombination. Since no essential functions have been found to be necessarily associated with the exogenous molecule it is beleived that this process faithfully reflects the natural recombination mechanisms of the cell. The DNA tumour virus paradigm has led to the development of a linear "replacement" model for integration by illegitimate recombination - model which predicts an excision of host DNA similar to the length of DNA inserted. RmI is a naturally occurring thermosensitive Py recombinant that contains an insertion (INS) of mouse cellular DNA. RmI should have the potential to integrate by any or all three of site-specific, illegitimate, or homologous recombination mechanisms. Firstly, RmI contains viral Py sequences known to integrate by illegitimate recombination. Secondly, RmI contains B2 and MT repetitive sequences both of which have 100,000 homologous copies dispersed throughout the rat genome any one of which could serve as a target for homologous integration. Finally the junctions between Ins and the Py sequences of RmI display features that suggest that RmI may be capable of site-specific recombination. My previous work on the integration of RmI into the genomes of rat cells showed that at the temperature non-permissive for replication, integration was non random with respect to the sequences of RmI. One region of RmI was underrepresented, whereas two regions, including the region of Ins containing the repetitive elements B2 and MT, were significantly overrepresented. Furthermore, the lengths of the integrated genomes tended importantly toward the unit length of RmI contrary to what was observed at the permissive temperature. In an effort to further define the integration mechanism and in particular to determine the nature of the viral-cellular junctions at the molecular level as well as to characterize the host site before and after integration, the cellular DNAs flanking 6 junctions from 3 transformed cell lines were cloned (4 in this study, 2 were previously cloned) and the sequences across 4 of these junctions were determined (one previously determined, 3 were determined here). The cloned DNAs were used to map the rearrangements of the cellular DNA caused by the integration of RmI. Furthermore, the fate of the rearranged intervening host sequences in one of the clones was determined. The results show that the overrepresentation of the repetitive sequences was due to their acting as hotspots for illegitimate recombination. Even though the host genome contained over 105 potential targets for homologous recombination we found no evidence for homologous integration, and conclude that it is not an efficient procedure in the case of RmI. Furthermore, contrary to previous reports which suggested that exogenous DNA integrated preferentially into repetitive sequences of the host, we found that the DNA flanking all but one junction was unique sequence DNA. Even though RmI was transfected in the presence of carrier DNA, the results presented here show that it integrated directly into the rat cellular DNA. This shows that transgenomes are not necessary intermediates for the integration of (circular) DNA even when transfected in the presence of carrier. The minimum lengths of the rearrangements caused to the host DNA by RmI's integration into two of the clones was determined to be 12 kbp for an insertion of 7 kbp and 55 kbp for an insertion of 6 kbp. It was further established that the intervening host DNA in the latter clone had been deleted. We show (for the first time by hybridization) that in the case of the third clone the DNAs flanking the insertion on either side were linked before integration. The rearrangement caused by RmI's integration is complex but can be partially explained by an inversion event of approximately 20 kbp for an insertion of 7 kbp. The length and nature of the rearrangements are incompatible with a linear insertion-replacement model, and are more easily explained by an integration mechanism in which the incoming exogenous DNA recombines with a looped structure of chromosomal DNA. Such a model, which takes into account recent findings on chromatin structure, is presented.