SummaryBorder fragments of the octopine Ti-plasmid were tested for their ability to restore tumorigenicity of an avirulent mutant carrying a deleted right border. It was found that neither introduction of left border fragments nor that of small right border fragments at the position of the deletion resulted in a complete restoration of oncogenicity. However, insertion of a larger right border fragment in the deletion mutant gave fully oncogenic strains. In the latter case sequences to the right side of the right border repeat were found to be responsible for a complete restoration of oncogenicity. Also a left border repeat inserted together with this enhancer sequence fully restored the oncogenicity of the deletion mutant. The enhancer-sequence on itself was not able to mediate the transfer of the T-region to the plant cell. Border fragments inserted in inverted orientation in the deletion mutant were able to mediate the transfer of the T-region to the plant cell, but at a reduced frequency.

We describe the use of plasmid rescue to facilitate studies on the behaviour of Ds and Ac elements in transgenic tomato plants. The rescue of Ds elements relies on the presence of a plasmid origin of replication and a marker gene selective in Escherichia coli within the element. The position within the gehome of modified Ds elements, rescued both before and after transposition, is assigned to the RFLP map of tomato. Alternatively to the rescue of Ds elements equipped with plasmid sequences, Ac elements are rescued by virtue of plasmid sequences flanking the element. In this way, the consequences of the presence of an (active) Ac element on the D N A structure at the original site can be studied in detail. Analysis of a library of Ac elements, rescued from the genome of a primary transformant, shows that Ac elements are, infrequently, involved in the formation of deletions. In one case the deletion refers to a 174 bp genomic D N A sequence immediately flanking Ac. In another case, a 1878 bp internal Ac sequence is deleted.

Fruit-specific expression of beta-glucuronidase (GUS) activity was produced in transgenic tomato plants when the GUS-coding region was flanked by 5' and 3' regions of the tomato 2A11 gene. Deletion studies on the 5' region revealed a number of strong regulatory elements involved in the proper expression of the 2A11 gene. A 4.0 kb and a 1.3 kb 5' region can confer high-level fruit-specific GUS expression, while a 1.8 kb 5' region produces no GUS activity in leaf or fruit tissue. Thus, a strong negative regulatory element is present in the region between 1324 bp and 1796 bp upstream of the 2A11 transcriptional start and a strong fruit-specific positive regulatory element is present more than 1.8 kb upstream of the transcriptional start site. The 1.8 kb promoter region can be activated by the upstream insertion of the CaMV 35S enhancer sequence, albeit not in a fruit-specific fashion. Substitution of the 3' region of the 2A11 gene with a different 3' region does not seem to affect GUS expression significantly, indicating a minor role, if any, for the 3' region in the fruit-specific expression of the 2A11 gene.

Introduction of a left or right synthetic border repeat together with the overdrive sequence in an octopine Ti-plasmid deletion mutant, lacking the right border, resulted in the complete restoration of the oncogenicity of the mutant strain. However introduction of a border repeat without the overdrive, only restored oncogenicity partially. The overdrive sequence turned out to be able to stimulate the synthetic border mediated T-region transfer, independent of its orientation and position relative to the border repeat. Furthermore the distance between border repeat and overdrive could be enlarged, without a loss of overdrive activity. Here we enlarged the distance between the two sequences up to 6714bp. These results were confirmed by estimating the amount of single stranded T-DNA molecules from induced agrobacteria, containing the various border constructs.

With the aim of developing new techniques for physical and functional genome analysis, we have introduced the Cre-lox site-specific recombination system into the cultivated tomato (Lycopersicon esculentum). Local transposition of a Ds(lox) transposable element from a T-DNA(lox) on the long arm of chromosome 6 was used to position pairs of lox sites on different closely linked loci. In vitro Cre-lox recombination between chromosomal lox sites and synthetic lox oligonucleotides cleaved the 750 Mb tomato genome with 34 bp specificity to release unique 65 kb and 130 kb fragments of chromosome 6. Parallel in vitro experiments on Saccharomyces cerevisiae chromosomes show the efficiency of cleavage to be 50% per chromosomal lox site at maximum. By expressing the Cre recombinase in tomato under control of a constitutive CaMV 35S promoter, efficient and specific somatic and germinal in planta inversion of the 130 kb fragment is demonstrated. The combined use of in vitro and in vivo recombination on genetically mapped lox sites will provide new possibilities for long range restriction mapping and in vivo manipulation of selected tomato genome segments.