Search

Cre/lox recombination in vivo has become an important tool to induce chromosomal rearrangements like deletions. Using a combination of Ds transposition and Cre/lox recombination in two independent experiments on chromosomes 6 and 7 of tomato, two sets of somatic deletions up to a size of 200 kb were obtained. The efficiency of somatic deletion decreased with increasing deletion size. The largest germinally transmitted deletion had a size of only 55 kb. The results show that Cre-mediated deletion in somatic cells is less efficient when the lox sites are separated over larger distances. A further drop of the deletion efficiency after germinal transmission of the larger deletions can be explained by the probable loss of genes that are of vital importance to gametophyte function. Plasmid rescue of an 8.4 kb circularised deleted DNA showed that the Cre-mediated deletion takes place in tomato as expected. Since the circular Cre-deleted DNA could only be PCR amplified in plant cells where the deletion was not complete, the double-stranded DNA circle is assumed to be instable.

In order to make the tomato genome more accessible for molecular analysis and gene cloning, we have produced 405 individual tomato (Lycopersicon esculentum) lines containing a characterized copy of pJasm13, a multifunctional T-DNA/modified Ds transposon element construct. Both the T-DNA and the Ds element in pJasm13 harbor a set of selectable marker genes to monitor excision and reintegration of Ds and additionally, target sequences for rare cutting restriction enzymes (I-PpoI, SfiI, NotI) and for site-specific recombinases (Cre, FLP, R). Blast analysis of flanking genomic sequences of 174 T-DNA inserts revealed homology to transcribed genes in 69 (40%), of which about half are known or putatively identified as genes and ESTs. The map position of 140 individual inserts was determined on the molecular genetic map of tomato. These inserts are distributed over the 12 chromosomes of tomato, allowing targeted and non-targeted transposon tagging, marking of closely linked genes of interest and induction of chromosomal rearrangements including translocations or creation of saturation-deletions or inversions within defined regions linked to the T-DNA insertion site. The different features of pJasm13 were successfully tested in tomato and Arabidopsis thaliana, thus providing a new tool for molecular/genetic dissection studies, including molecular and physical mapping, mutation analysis and cloning strategies in tomato and potentially, in other plants as well.

The cre recombinase gene was stably introduced and expressed in tomato, petunia and Nicotiana tabacum. Some plants expressing the cre gene driven by a CaMV 35S promoter displayed growth retardation and a distinct pattern of chlorosis in their leaves. Although no direct relation can be proven between the phenotype and cre expression, aberrant phenotypes always co-segregate with the transgene, which strongly suggests a correlation. The severity of the phenotype does not correlate with the level of steady-state mRNA in mature leaves, but with the timing of cre expression during organogenesis. The early onset of cre expression in tomato is correlated with a more severe phenotype and with higher germinal transmission frequencies of site-specific deletions. No aberrant phenotype was observed when a tissue-specific phaseolin promoter was used to drive the cre gene. The data suggest that for the application of recombinases in plants, expression is best limited to specific tissues and a short time frame.

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.

The concept of gene identification and cloning using insertional mutagenesis is well established. Many genes have been isolated using T-DNA transformation or transposable elements. Maize transposable elements have been introduced into heterologous plant species for tagging experiments. The behaviour of these elements in heterologous hosts shows many similarities with transposon behaviour in Zea mays. Site-specific recombination systems from lower organisms have also been shown to function efficiently in plant cells. Combining transposon and site-specific recombination systems in plants would create the possibility to induce chromosomal deletions. This 'transposition-deletion' system could allow the screening of large segments of the genome for interesting genes and may also permit the cloning of the DNA corresponding to the deleted material by the same site-specific recombination reaction in vitro. This methodology may provide a unique means to construct libraries of large DNA clones derived from defined parts of the genome, the phenotypic contribution of which is displayed by the mutant carrying the deletion.

Studies on transposable elements of the Ac family have led to different models for excision gap repair in either plants or Drosophila. Excision products generated by the plant transposable elements Ac and Tam3 imply a more or less straightforward ligation of broken ends; excision products of the Drosophila P element indicate the involvement of 'double-strand break' (DSB) repair. Recent findings that excision products of Ac and Tam3 can also contain traces of the element ends indicate, however, that DSB repair might be an alternative repair mechanism in plants. A functional DSB repair mechanism in plants can also be deduced from the observed rapid increases of Ac copy number during plant development and from the involvement of Ac in the generation of internal Ac deletions. On the other hand, alternative repair mechanisms may also be functional in Drosophila, because some of the 'footprints' generated upon P excision can be explained by a mechanism that has been postulated for excision gap repair in plants. It is concluded that plants and Drosophila can use similar repair mechanisms, but that the predominance of a certain repair mechanism is determined by the host.

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 genome 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 DNA 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 DNA sequence immediately flanking Ac. In another case, a 1878 bp internal Ac sequence is deleted.

To study regulation of the (Ds) transposition process in heterologous plant species, the transposase gene of Ac was fused to several promoters that are active late during plant development. These promoters are the flower-specific chalcone synthase A promoter (CHS A), the anther-specific chalcone isomerase B promoter CHI B and the pollen-specific chalcone isomerase A2 promoter CHI A2. The modified transposase genes were introduced into a tobacco tester plant. This plant contains Ds stably inserted within the leader sequence of the hygromycin resistance (HPT II) gene. As confirmed with positive control elements, excision of Ds leads to the restoration of a functional HPT II gene and to a hygromycin resistant phenotype. No hygromycin resistance was observed in negative control experiments with Ac derivatives lacking 5' regulatory sequences. Although transactivation of Ds was observed after the introduction of transposase gene fusions in calli, excision in regenerated plants was observed only for the CHS A- or CHI B-transposase gene fusions. With these modified transposase genes, somatic excision frequencies were increased (68%) and decreased (22%), respectively, compared to the situation with the Ac element itself (38%). The shifts in transactivation frequencies were not associated with significant differences in the frequencies of germinally transmitted excision events (approximately 5%). The relative somatic stability of Ds insertions bearing the CHI B-transposase gene fusion suggests the usefulness of this activator element for transposon tagging experiments.

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.

Agrobacterium strains harbouring the T-region and the virulence-region of the Ti plasmid on separate replicons still display efficient T-DNA transfer to plants. Based on this binary vector strategy we have constructed T-region derived gene vectors for the introduction of foreign DNA into plants. The vectors constructed can replicate in E. coli, thus the genetic manipulations with them can be performed with E. coli as a host. They can be transferred to Agrobacterium as a cointegrate with the wide host range plasmid R772. Their T-regions are transferred to plant cells from Agrobacterium strains conferring virulence functions.The plasmid pRAL 3940 reported here is 11.5 kb large, contains a marker to identify transformed plant cells and unique restriction sites for direct cloning of passenger DNA, flanked by the left- and right-hand border fragments of the T-region (including the 25 bp border repeats). The plasmid is free of onc-genes. Therefore, is does not confer tumorigenic traits on the transformed plant cells and mature, fertile plants can thus be regenerated from them.

T-region transfer from wild-type Agrobacterium strains is thought to be an orientated process, starting at the right border repeat and terminating at the left border repeat of the T-region. Here we demonstrate that a right border repeat in the inverted orientation relative to the onc-genes can also mediate transfer of the T-region to the plant cell, although with lower efficiency as a border repeat in the native orientation. Transfer mediated by an inverted right border repeat is stimulated by the presence of the T-region transfer enhancer. Similar single stranded molecules, comprising the bottom strand of the T-DNA, were isolated from acetosyringone induced bacteria, irrespective of the orientation of the right border. These findings show that border repeats work bidirectionally to some extent.

Left- and right-border repeats, which surround the T-region, contain two conserved regions separated by 5 bp that are not conserved. At the onset of T-DNA processing virD-encoded proteins introduce a nick in the largest of these conserved regions (12 bp) at a specific position in the bottom strand between a guanine and thymine nucleotide [2, 33]. In this paper we describe the effect of several site-directed mutations in the right-border repeat on tumorigenicity of Agrobacterium in plants. Our data show that mutations introduced directly around the nick site do not seriously affect the tumorigenicity of Agrobacterium, whereas mutations in the right part of this 12 bp conserved region do so. Furthermore, it appeared that the second conserved region (5 bp) is also essential for border activity and that the distance between the two conserved regions is important to obtain optimal border activity.

Border 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.

The successful transfer of the Ti plasmid T region to the plant cell is mediated by its 24 bp border repeats. Processing of the T-region prior to transfer to the plant cell is started at the right border repeat and is stimulated by a transfer enhancer sequence called "overdrive". Left and right border repeats differ somewhat in nucleotide sequence; moreover, the repeats of different Ti and Ri plasmids are slightly different. Our data indicate that these differences do not have a significant influence on border activity. However, the overdrive sequence is essential for the efficient transfer of a T region via an octopine transfer system. Our data suggest that an overdrive sequence must also be present next to the right border repeats of the nopaline Ti plasmid and the agropine of octopine and nopaline Ti plasmids express some differences in T-DNA processing activities. of cotopine and nopaline Ti plasmids express some differences in T-DNA processing activities.Furthermore, we demonstrate that certain pseudo border repeats, sequences that resemble the native 24 bp border repeat and naturally occur within the octopine Ti plasmid T-region, are able to mediate T region transfer to the plant cell, albeit with much reduced efficiency as compared to wild-type border repeats.

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.