Integration of Agrobacterium T-DNA into a tobacco chromosome: Possible involvement of DNA homology between T-DNA and plant DNA

Summary We established tobacco tumour cell lines from crown galls induced by Agrobacterium. Restriction fragments containing T-DNA/plant DNA junctions were cloned from one of the cell lines, which has a single copy of the T-DNA in a unique region of its genome. We also isolated a DNA fragment that contained the integration target site from nontransformed tobacco cells. Nucleotide sequence analyses showed that the right and left breakpoints of the T-DNA mapped ca. 7.3 kb internal to the right 25 by border and ca. 350 by internal to the left border respectively. When the nucleotide sequences around these breakpoints were compared with the sequence of the target, significant homology was seen between the region adjacent to the integration target site and both external regions of the T-DNA breakpoints. In addition, a short stretch of plant DNA in the vicinity of the integration site was deleted. This deletion seems to have been promoted by homologous recombination between short repeated sequences that were present on both sides of the deleted stretch. Minor rearrangements, which included base substitutions, insertions and deletions, also took place around the integration site in the plant DNA. These results, together with previously reported results showing that in some cases sequences homologous to those in T-DNA are present in plant DNA regions adjacent to left recombinational junctions, indicate that sequence homology between the incoming T-DNA and the plant chromosomal DNA has an important function in T-DNA integration. The homology may promote close association of both termini of a T-DNA molecule on a target sequence; then TDNA may in some cases be integrated by a mechanism at least in part analogous to homologous recombination.Shogo Matsumoto is on leave from Biochemical Research Institute, Nippon Menard Cosmetic Co., Ltd, Ogaki, Gifu-ken 503, Japan     

Features of two hepatitis B virus (HBV) DNA integrations suggest mechanisms of HBV integration.

Two integrated hepatitis B virus (HBV) DNA molecules were cloned from two primary hepatocellular carcinomas each containing only a single integration. One integration (C3) contained a single linear segment of HBV DNA, and the other integration (C4) contained a large inverted duplication of viral DNA at the site of a chromosome translocation (O. Hino, T.B. Shows, and C.E. Rogler, Proc. Natl. Acad. Sci. USA 83:8338-8342, 1986). Sequence analysis of the virus-cell junctions of C3 placed the left virus-cell junction at nucleotide 1824, which is at the 5' end of the directly repeated DR1 sequence and is 6 base pairs from the 3' end of the long (L) negative strand. The right virus-cell junction was at nucleotide 1762 in a region of viral DNA (within the cohesive overlap) which shared 5-base-pair homology with cellular DNA. Sequence analysis of the normal cellular DNA across the integration site showed that 11 base pairs of cellular DNA were deleted at the site of integration. On the basis of this analysis, we suggest a mechanism for integration of the viral DNA molecule which involves strand invasion of the 3' end of the L negative strand of an open circular or linear HBV DNA molecule (at the DR1 sequence) and base pairing of the opposite end of the molecule with cellular DNA, accompanied by the deletion of 11 base pairs of cellular DNA during the double recombination event. Sequencing across the inverted duplication of HBV DNA in clone C4 located one side of the inversion at nucleotide 1820, which is 2 base pairs from the 3' end of the L negative strand. Both this sequence and the left virus-cell junction of C3 are within the 9-nucleotide terminally redundant region of the HBV L negative strand DNA. We suggest that the terminal redundancy is a preferred topoisomerase I nicking region because of both its base sequence and forked structure. Such nicking would lead to integration and rearrangement of HBV molecules within the terminal redundancy, as we have observed in both our clones.     

Sequence requirements for DNA rearrangements induced by the terminal repeat of herpes simplex virus type 1 KOS DNA.

We investigated the sequence requirements for the site-specific DNA cleavages and recombinational genome isomerization events driven by the terminal repeat or a sequence of herpes simplex virus type 1 KOS DNA by inserting a series of mutated a sequences into the thymidine kinase locus in the intact viral genome. Our results indicate that sequences located at both extremities of the a sequence contribute to these events. Deletions entering from the Ub side of the a sequence progressively reduced the frequency of DNA rearrangements, and further deletion of the internal DR2 repeat array had an additional inhibitory effect. This deletion series allowed us to map the pac1 site-specific DNA cleavage signal specifying the S-terminal cleavage to a sequence that is conserved among herpesvirus genomes. Constructs lacking this signal were unable to directly specify the S-terminal cleavage event but retained a reduced ability to give rise to S termini following recombination with intact a sequences. Deletions entering from the Uc side demonstrated that the copy of direct repeat 1 located adjacent to the Uc region plays an important role in the DNA rearrangements induced by the a sequence: mutants lacking this sequence displayed a reduced frequency of novel terminal and recombinational inversion fragments, and further deletions of the Uc region had a relatively minor additional effect. By using a construct in which site-specific cleavage was directed to heterologous DNA sequences, we found that the recombination events leading to genome segment inversion did not occur at the sites of DNA cleavage used by the cleavage-packaging machinery. This observation, coupled with the finding that completely nonoverlapping portions of the a sequence retained detectable recombinational activity, suggests that inter-a recombination does not occur by cleavage-ligation at a single specific site in herpes simplex virus type 1 strain KOS. The mutational sensitivity of the extremities of the a sequence leads us to hypothesize that the site-specific DNA breaks induced by the cleavage-packaging system stimulate the initiation of recombination.     

The extent of DNA sequence required for a functional bacterial attachment site of phage lambda.

We have investigated the extent of DNA sequence required to form a bacterial attachment site (attB) that functions in bacteriophage lambda integration. A DNA fragment carrying attB of Escherichia coli was trimmed, recloned and tested for recombination proficiency. We found that the common core sequence plus the adjoining 4-bp sequences of both the B and B' arms are required for full activity, while plasmids with an even shorter attB sequence retain some capacity to function as attB in vivo. We also found that the nonspecific DNA that is joined to the required attachment site sequence does not significantly influence the rate of the recombination reaction.     

Structural and biological analysis of integrated polyoma virus DNA and its adjacent host sequences cloned from transformed rat cells.

EcoRI fragments containing integrated viral and adjacent host sequences were cloned from two polyoma virus-transformed cell lines (7axT and 7axB) which each contain a single insert of polyoma virus DNA. Cloned DNA fragments which contained a complete coding capacity for the polyoma virus middle and small T-antigens were capable of transforming rat cells in vitro. Analysis of the flanking sequences indicated that rat DNA had been reorganized or deleted at the sites of polyoma virus integration, but none of the hallmarks of retroviral integration, such as the duplication of host DNA, were apparent. There was no obvious similarity of DNA sequences in the four virus-host joins. In one case the virus-host junction sequence predicted the virus-host fusion protein which was detected in the transformed cell line. DNA homologous to the flanking sequences of three out of four of the joins was present in single copy in untransformed cells. One copy of the flanking host sequences existed in an unaltered form in the two transformed cell lines, indicating that a haploid copy of the viral transforming sequences is sufficient to maintain transformation. The flanking sequences from one cell line were further used as a probe to isolate a target site (unoccupied site) for polyoma virus integration from uninfected cellular DNA. The restriction map of this DNA was in agreement with that of the flanking sequences, but the sequence of the unoccupied site indicated that viral integration did not involve a simple recombination event between viral and cellular sequences. Instead, sequence rearrangements or alterations occurred immediately adjacent to the viral insert, possibly as a consequence of the integration of viral DNA.     

Characterization of unintegrated retroviral DNA with long terminal repeat-associated cell-derived inserts.

We have used a replication-competent shuttle vector based on the genome of Rous sarcoma virus to characterize genomic rearrangements that occur during retrovirus replication. The strategy involved cloning circular DNA that was generated during an acute infection. While analyzing a class of retroviral DNA clones that are greater than full length, we found several clones which had acquired nonviral inserts in positions adjacent to the long terminal repeats (LTRs). There appear to be two distinct mechanisms leading to the incorporation of cellular sequences into these clones. Three of the molecules contain a cell-derived insert at the circle junction site between two LTR units. Two of these molecules appear to be the results of abortive integration attempts, because of which, in each case, one of the LTRs is missing 2 bases at its junction with the cell-derived insert. In the third clone, pNO220, the cellular sequences are flanked by an inappropriately placed copy of the tRNA primer-binding site on one side and a partial copy of the U3 sequence as part of the LTR on the other side. A fourth molecule we characterized, pMD96, has a single LTR with a U5-bounded deletion of viral sequences spanning gag and pol, with cell-derived sequences inserted at the site of the deletion; its origin may be related mechanistically to pNO220. Sequence analysis indicates that all of the cellular inserts were derived from the cell line used for the acute infection rather than from sequences carried into the cell as part of the virus particle. Northern (RNA) analysis of cellular RNA demonstrated that the cell-derived sequences of two clones, pNO220 and pMD96, were expressed as polyadenylated RNA in uninfected cells. One mechanism for the joining of viral and cellular sequences suggested by the structures of pNO220 and pMD96 is recombination occurring during viral DNA synthesis, with cellular RNA serving as the template for the acquisition of cellular sequences.     

Role of arginine-43 and arginine-69 of the Hin recombinase catalytic domain in the binding of Hin to the hix DNA recombination sites

The Hin recombinase mediates the site-specific inversion of a segment of the Salmonella chromosome between two flanking 26 bp hix DNA recombination sites. Mutations in two amino acid residues, R43 and R69 of the catalytic domain of the Hin recombinase, were identified that can compensate for loss of binding resulting from elimination of certain major and minor groove contacts within the hix recombination sites. With one exception, the R43 and R69 mutants were also able to bind a hix sequence with an additional 4 bp added to the centre of the site, unlike wild-type Hin. Purified Hin mutants R43H and R69C had both partial cleavage and inversion activities in vitro while mutants R43L, R43C, R69S, and R69P had no detectable cleavage and inversion activities. These data support a model in which the catalytic domain plays a role in DNA-binding specificity, and suggest that the arginine residues at positions 43 and 69 function to position the Hin recombinase on the DNA for a step in the recombination reaction which occurs either at and/or prior to DNA cleavage.     

DNA rearrangement associated with the integration of T-DNA in tobacco: an example for multiple duplications of DNA around the integration target

Transferred DNA (T-DNA) of the tumor-inducing (Ti) plasmid is transferred from Agrobacterium tumefaciens to plant cells and is stably integrated into the plant nuclear genome. By the inverse polymerase chain reaction DNA fragments were amplified that contained the T-DNA/plant DNA junctions from the total DNA of a transgenic tobacco plant that had a single copy of the T-DNA in a repetitive region of its genome. A DNA fragment containing the target site was amplified from the total DNA of non-transformed tobacco by the polymerase chain reaction using high-stringency conditions. Comparison of the nucleotide sequence of the target site with those of the T-DNA/plant DNA junctions revealed that various duplications of short stretches of nucleotide sequences around the target and in the incoming T-DNA had accompanied the integration of the T-DNA. A deletion of 16 bp at the target site was also found and the target site was similar, in terms of nucleotide sequence, to regions around the breakpoints of the T-DNA. This finding provides a clear example of the occurrence of complex rearrangements during the integration of T-DNA.     

Structure of an inverted duplication formed as a first step in a gene amplification event: implications for a model of gene amplification.

Inverted duplications have been observed to be a common feature of gene amplification in mammalian cells and appear to be generated as a primary event in the amplification process (Ford et al., 1985; Ford and Fried, 1986). The structural features of the amplified inverted duplication, containing the polyoma virus oncogene middle T-antigen, were analysed in transformed 3B rat cells. No unusual sequences such as transposition elements were detected at the site of the inversion. The inversion was generated by a simple illegitimate recombination event in which only a single nucleotide directly at the point of the inversion cannot be accounted for from the sequence of the two parental strands. Possible structural (hairpin formation) and sequence (rich AT) features may have been involved in the illegitimate recombination event at the inversion join. In the cellular DNA near one of its joins with polyoma virus DNA an unusual sequence of 198 bp composed of 99 consecutive purine-pyrimidine pairs has been detected. A model for the generation of amplified DNA containing inverted duplications is proposed.     

Single-stranded DNA as a recombination substrate in plants as assessed by stable and transient recombination assays.

Two separate assays, one that requires stable integration of recombination products and one that does not, were employed to elucidate the role of single-stranded DNA in extrachromosomal homologous recombination in Nicotiana tabacum. Both assays revealed that single-stranded DNA in linear and in circular forms was an efficient substrate for recombination, provided that the cotransformed recombination substrates were of complementary sequence, so that direct annealing was possible. Recombination was inefficient when both single-stranded recombination partners contained homologous regions of identical sequence and generation of a double-stranded DNA was required prior to heteroduplex formation. These results indicate that direct annealing of single strands is an important initial step for intermolecular recombination in tobacco cells. Annealed cotransformed single-stranded molecules yielded intermediates that could be further processed by either continuous or discontinuous second-strand synthesis. The type of intermediate had no influence on the recombination efficiency. Double-stranded circles were unable to recombine efficiently either with each other or with single-stranded DNA. Our results suggest that a helicase activity is involved in the initial steps of double-stranded DNA recombination which unwinds duplex molecules at the site of double-strand breaks.     

Requirement for double-strand breaks but not for specific DNA sequences in herpes simplex virus type 1 genome isomerization events.

Herpes simplex virus type 1 (HSV-1) genome isomerization occurs as a result of DNA replication-mediated homologous recombination between several sets of inverted repeat sequences present in the viral DNA. The frequency with which this recombination occurs has been demonstrated to be dependent upon DNA homology length rather than specific sequences. However, the smallest of the viral inverted repeats, the alpha sequence, has been shown to function as a recombinational hot spot, leading to speculation that this sequence may represent a specific element through which genome isomerization is mediated. To investigate this apparent paradox, a quantitative transient recombination assay system was developed and used to examine the recombinogenic properties of a panel of alpha sequence mutants. This analysis revealed that the presence of both the pac1 and pac2 elements was both necessary and sufficient for the induction of high-frequency recombination events by the alpha sequence. However, it was the double-strand break promoted by pac1 and pac2 during cleavage and packaging at the alpha sequence, and not the DNA sequences of the elements themselves, which appeared to be critical for recombination. This was illustrated (i) by the inability of the same pac1 and pac2 sequences to mediate inversion events in cells infected with an HSV-1 mutant which was competent for DNA replication-dependent recombination but defective for the cleavage and packaging process and (ii) by the ability of double-strand breaks generated in non-HSV-1 DNA by an in vivo-expressed restriction endonuclease to significantly stimulate the initiation of recombination events in virus-infected cells. Thus, the alpha sequence appears to act as a hot spot for homologous recombination simply because it happens to coincide with the site of the double-strand break which is generated during the cleavage and packaging process, not because it contains discrete sequences which are required for this activity. However, it was found that this enhanced recombinogenicity disappeared when the element was flanked by regions of extensive sequence homology, particularly that of the large inverted repeats which flank the alpha sequence at its natural site in the HSV-1 genome. These findings are consistent with a model for HSV-1 genome isomerization in which recombination is initiated primarily by multiple random double-strand breaks which arise during DNA replication across the inverted repeats of the genome, rather than by a single specific break which occurs at the alpha sequence during the cleavage and packaging process.     

Intermediates in Hin-mediated DNA inversion: a role for Fis and the recombinational enhancer in the strand exchange reaction.

The site-specific inversion reaction controlling flagellin synthesis in Salmonella involves the function of three proteins: Hin, Fis and HU. The DNA substrate must be supercoiled and contain a recombinational enhancer sequence in addition to the two recombination sites. Using mutant substrates or modified reaction conditions, large amounts of complexes can be generated which are recognized by double-stranded breaks within both recombination sites upon quenching. The cleaved molecules contain 2-bp staggered cuts within the central dinucleotide of the recombination site. Hin is covalently associated with the 5' end while the protruding 3' end contains a free hydoxyl. We demonstrate that complexes generated in the presence of an active enhancer are intermediates that have advanced past the major rate limiting step(s) of the reaction. In the absence of a functional enhancer, Hin is also able to assemble and catalyze site-specific cleavages within the two recombination sites. However, these complexes are kinetically distinct from the complexes assembled with a functional enhancer and cannot generate inversion without an active enhancer. The results suggest that strand exchange leading to inversion is mediated by double-stranded cleavage of DNA at both recombination sites followed by the rotation of strands to position the DNA into the recombinant configuration. The role of the enhancer and DNA supercoiling in these reactions is discussed.     

Recombination events during integration of transfected DNA into normal human cells.

The mechanisms of recombination responsible for random integration of transfected DNA into the genome of normal human cells have been investigated by analysis of plasmid-cell DNA junctions. Cell clones containing integrated plasmid sequences were selected by morphological transformation of primary human fibroblasts after transfection with a plasmid containing simian virus 40 sequences. Nucleotide sequence analysis of the plasmid-cell DNA junctions was performed on cloned DNA fragments containing the integration sites from two of these cell clones. Polymerase chain reaction was then performed with human cell DNA from primary fibroblasts to isolate the cell DNA from the same sites before plasmid integration. Comparison of the sequences at the plasmid-cell DNA junctions with those of both the original plasmid and the cell DNA demonstrated short sequence similarities and additional nucleotides, typical of nonhomologous recombination. Evidence of short deletions in the cell DNA at the plasmid integration sites suggests that integration occurred by a mechanism similar to that used for repair of spontaneous or gamma ray-induced strand breaks. Plasmid integration occurred within nonrepetitive cell DNA with no major rearrangements, although rearrangements of the cell DNA at the integration site occurred in one of the clones after integration.     

Microhomologies between T-DNA ends and target sites often occur in inverted orientation and may be responsible for the high frequency of T-DNA-associated inversions

Sequence analysis of left and right border integration sites of independent, single-copy T-DNA inserts in Arabidopsis thaliana revealed three previously unrecognized concomitants of T-DNA integration. First, genomic pre-insertion sites shared sequence similarity not only with the T-DNA left and right border regions, as was previously reported, but also at high frequency with the inverted complement of the T-DNA right border region. Second, palindromic sequences were frequently found to overlap or lie adjacent to genomic target sites, suggesting a high recombinogenic potential for palindromic elements during T-DNA integration and a possible role during the primary contact between the T-DNA and the target DNA. Third, “filler” DNA sequences between genomic pre-insertion site DNA and T-DNA often derive from sequences in the T-DNA left and right border regions that are clustered around palindromic sequences in these T-DNA regions, suggesting that these palindromic elements are “hot spots” for filler DNA formation. The discovery of inverted sequence similarities at the right border suggests a previously unrecognized mode of T-DNA integration that involves heteroduplex formation at both T-DNA borders and with opposite strands of the target DNA. Scanning for sequence similarities in both direct and inverted orientation may increase the probability and/or effectiveness of anchoring the T-DNA to the target DNA. Variations on this scheme may also account for inversion events at the target site of T-DNA integration and inverted T-DNA repeat formation, common sequence organization patterns associated with T-DNA integration. Electronic Supplementary Material Supplementary material is available in the online version of this article at and is accessible for authorized users.     

The structures of integration sites in transgenic rice

Extensive genomic sequencing and sequence motif analysis have been conducted over the integration sites of two transgenic rice plants, #478 and #559, carrying the luciferase gene and/or hygromycin phosphotransferase gene. The transgenes reside in a region with inverted structure and a large duplication of rice genome over 2 kb. Integration was found at the AT-rich region and/or at the repetitive sequence region, including a SAR-like structure, retrotransposon and telomere repeats. The presence of a patch of sequence homology between plasmid and target DNA, and a small region of duplication involving the target DNA around the recombination site, implicated illegitimate recombination in the process of gene integration. Massive rearrangement of genomic DNA including deletion or translocation was also observed at the integration site and the flanking region of the transgene. The recognition sites of DNA topoisomerases I or II were observed in the rearranged sequences. Since only three junctions of transgenic rice were implicated in the illegitimate recombination and extensive rearrangement of the rice genome, rice protoplasts may be active in this process.     

In Vitro HIV-1 LTR Integration into T-Cell Activation Gene CD27 Segment and the Decoy Effect of Modified-Sequence DNA

Integration into the host genome is an essential step in the HIV-1 life cycle. However, the host genome sequence that is favored by HIV-1 during integration has never been documented. Here, we report that CD27, a T cell activation gene, includes a sequence that is a target for in vitro HIV-1 cDNA integration. This sequence has a high affinity for integrase, and the target nucleotides responsible for this higher affinity were identified using a crystal microbalance assay. In experiments involving a segment of the CD27 gene, integration converged in the target nucleotides and flanking sequence DNA, indicating that integration is probably dependent upon the secondary structure of the substrate DNA. Notably, decoy modified CD27 sequence DNAs in which the target nucleotides were replaced suppressed integration when accompanying the original CD27 sequence DNA. Our identified CD27 sequence DNA is useful for investigating the biochemistry of integrase and for in vitro assessment of integrase-binding inhibitors.     

Long terminal repeat of murine retroviral DNAs: sequence analysis, host-proviral junctions, and preintegration site.

The nucleotide sequence of the long terminal repeat (LTR) of three murine retroviral DNAs has been determined. The data indicate that the U5 region (sequences originating from the 5' end of the genome) of various LTRs is more conserved than the U3 region (sequences from the 3' end of the genome). The location and sequence of the control elements such as the 5' cap, "TATA-like" sequences, "CCAAT-box," and presumptive polyadenylic acid addition signal AATAAA in the various LTRs are nearly identical. Some murine retroviral DNAs contain a duplication of sequences within the LTR ranging in size from 58 to 100 base pairs. A variant of molecularly cloned Moloney murine sarcoma virus DNA in which one of the two LTRs integrated into the viral DNA was also analyzed. A 4-base-pair duplication was generated at the site of integration of LTR in the viral DNA. The host-viral junction of two molecularly cloned AKR-murine leukemia virus DNAs (clones 623 and 614) was determined. In the case of AKR-623 DNA, a 3- or 4-base-pair direct repeat of cellular sequences flanking the viral DNA was observed. However, AKR-614 DNA contained a 5-base-pair repeat of cellular sequences. The nucleotide sequence of the preintegration site of AKR-623 DNA revealed that the cellular sequences duplicated during integration are present only once. Finally, a striking homology between the sequences flanking the preintegration site and viral LTRs was observed.