CaMV Promoter is A Recombination Hotspot - No Transgenic Plant Containing CaMV Promoter Should be Released

Prepared by Angela Ryan, Molecular Biologist, Open University 16jul99

Lay Summary
A recent study of transgenic rice carried out at the John Innes Institute [1] supports previous evidence that there is a 'recombination hotspot' in the CaMV 35S promoter. A recombination hotspot is a site prone to recombination, ie, breaking and joining with other DNA. Furthermore, some of the recombination events are 'illegitimate' or nonhomologous, and do not require substantial similarity in nucleic acid base sequence.

Implications
The results show that the CaMV promoter is very likely to recombine with other DNA in the host genome, including dormant viral DNA, as well as with other viruses in the host cell. Transgenic lines containing CaMV promoters, which includes practically all that have been released, are therefore prone to instability due to rearrangements, and also have the potential to create new viruses or other invasive genetic elements.

Such elements cannot be contained or controlled once they have entered the wider environment. It is now indisputable that recombination events will take place at the CaMV promoter in the current generation of transgenic plants. The continued release of such transgenic plants is unwarranted especially in the light of the new findings.

Technical details
Twelve representative transgenic rice lines were analyzed, carrying a range of transforming plasmid rearrangements, which predominantly reflected micro-homology mediated illigitimate recombination involving short complementary patches at the recombining ends. Direct end-ligation (ie, joining of ends), in the absense of homology between recombining molecules, was also observed but occurred less frequently. Filler DNA was found at some of the junctions and short purine rich tracts were also found either at the junction or in the immediate flanking regions. Furthermore, putative DNA topoisomerase I binding sites were found in clusters around the junction.

Links between DNA double strand break repair (DSBR), illegitimate recombination and plasmid DNA integration have previously been established and involve sequences with either microhomology or no homology. This study reveals that there are similarities between recombination junctions generated by various transformation methods and this strongly suggests that the underlying mechanisms controlling plasmid rearrangement and transgene integration in plants are likely
to be the same.

Intergration of foreign DNA has been studied in detail in animal genomes and it appears that large amounts of DNA ends up stimulating the production of DNA ligase, which in turn promotes illegitimate recombination. A wound response is elicited in both Agrobacterium-mediated DNA delivery and direct physical DNA transfer into plant cells. This involves the activation of nucleases and DNA repair enzymes which maintain the integrity of the host genome. When unorthodox
substrates are present, illegitimate recombinations can lead to large scale genome rearrangements and the integration of exogenous DNA. Any exogenous DNA entering the cell is therefore exposed to breakdown and repair enzymes, resulting in some rearrangement and/or incorporation of it into the recipient genome. DSBR is the predominant mechanism of illegitimate recombination in higher eukaryotes, probably due to the large genome size preventing homology searching and also the higher order chromatin structure holding broken DNA ends in close proximity.

Although different regions of transforming plasmid were involved in plasmid-plasmid recombination, a 19 bp palindromic sequence, including the TATA box of the CaMV 35S promoter acted as a recombination hotspot, ie, a  hotspot for breaking and joining up with other DNA. Furthermore, the palindrome and surrounding DNA sequence were found to possess a number of characteristics common to known recombination hotspots. The purine-rich half of the palindrominc sequence was specifically involved at the recombination junctions. AT-rich sequences cause isotropic DNA bending and influence DNA melting and have been shown to contain S/MAR motifs (Sawasaki et al 1998) which intrinsically harbor curved DNA. There is a short tract of alternating purine-pyrimidine residues situated 50 bp upstream. Such sequences are known to adopt a Z-DNA conformation which in turn is known to influence transcription and recombination . These sequences are also known to bind DNA topoisomerase II which is involved in the resolution of recombination intermediates. In addition, the 3' end of the CaMV promoter was found to have structure and sequence similarity to the petunia transformation booster sequence which is shown to increase plant transformation efficiency, most likely by stimulating recombination. Other similar structures were found in recombinogenic regions of SV40 DNA and HeLa cells. Furthermore the 25 bp border repeats of T-DNA shows a remarkable similarity to the recombination hotspot of the CaMV promoter: There is an 11 bp palindromic sequence involving a TATA box-like structure in the right border and the left border has a short purine-rich sequence in the center. This study predicts that these two regions of T-DNA could be involved in rearrangements and indeed certain crossover
events have been previously documented.

The recombination hotspot described in the CaMV 35S promoter is found within the highly recombinogenic region of the full-length CaMV RNA and this study shows that recombination events can occur in this region even in the absense of viral enzymes and other cis-acting elements. It was shown that in CaMV RNA the recombination events were clustered around the 35S RNA transcription initiation site. This site is believed to be involved in recombination during
reverse transcriptase-mediated virus replication. A template switch at the 5' end of the RNA is induced by the 19 S RNA terminal repeat. However, in this study concerning the 423 bp fragment of the CaMV promoter, recombinogenic activity was maintained in the absense of reverse transcriptase and the remainder of the virus genome. These results prove that the plant cellular machinery alone is sufficient to recognise and act on these viral sequences.

In one of the transgenic rice lines the junction included the insertion of a 23 bp fragment of filler DNA and the presense of direct repeats (5'TCCGG 3') flanking the insert, suggesting one of two possible mechanisms. The synthesis of untemplated nucleotides by illegitimate recombination between the two ends representing short tails of imperfect complementarity. Alternatively, the insertion may represent a transposition event whereby the presense of staggered breaks in a target DNA molecule may have acted as a substrate for the transposase or integrase encoded by an endogenous plant transposable element. Insertions ranging from 2 bp to 1.2 kb were found in another study in nearly 30% of the plasmid junctions analyzed. This so called filler DNA was sometimes genomic in origin, sometimes it appeared to have been derived from the transforming plasmid and in other cases the origin was unknown. The entire insertion could itself be defined as filler DNA or captured DNA and the possible involvement of transposase in the generation of plasmid-plasmid junctions exemplifies a discrete form of illegitimate recombination characterised by the use of incorrect substrates by various DNA processing enzymes. Such rearrangements have been seen frequently with transposases and integrases, and with the enzymes that catalyze site-specific recombination (e.g. Cre recombinase, l integrase and Hin invertase).

Reference

1. Kohli, A. 1999. Molecular characterization of transforming plasmid rearrangement in transgenic rice reveals a recombination hotsport in the CaMV promoter and confirms the predominace of microhomology mediated recombination. The Plant Journal 17(6), pp 591-601.