Structural effect of donor DNA on the initiation of recombination for double strand break repair in human nuclear extracts.

The effect of the structure of donor DNA molecules on the initiation of recombination for double strand break repair in human nuclear extracts, was investigated here. A unique double strand break was introduced into M13 duplex derivatives by digestion with restriction enzymes. After coincubation of the cleaved DNA in human nuclear extracts, with a plasmid containing M13 sequences spanning the break, double strand break repair was estimated by the plating efficiency in JM109 (RecA1) bacteria. We first confirm that a short heterologous insert (8bp) close to the break on the recipient cleaved M13 DNA inhibits recombination with circular as well as with linear donor molecules. The results indicate that, with these substrates, recombination is initiated at the level of the break, requires uninterrupted homology on both sides of the break, and is associated with a decreasing gradient of gene conversion. When the heterologous insertion is located on the plasmid donor DNA, similar results are obtained with a circular donor DNA. In contrast, with a linear donor molecule, bearing the insert, homology requirements, in the region of the break in M13 DNA, are abolished. This last result suggests that recombination could be initiated at the extremities of the linear donor DNA.     

Recombination of homologous DNA fragments transfected into mammalian cells occurs predominantly by terminal pairing.

The mechanism by which double-strand cleavages stimulate the joining of plasmid DNA fragments introduced into cultured mammalian cells was investigated by cotransfecting pairs of plasmids encoding deletion mutations in a dominant selectable gene into LMtk- cells. Plasmid recombination substrates were produced by creating deletions of different sizes within the neo coding region of the pSV2neo plasmid. Complementing pairs of deleted plasmid DNAs were linearized at specific unique sites before cotransfection into mouse LMtk- cells by the calcium phosphate precipitation method. Cleaving one donor plasmid produced a 4- to 10-fold stimulation in the production of colonies able to survive in medium containing G-418. The linearization of the second plasmid further increased the efficiency by another factor of 6 to 15 when the cut was made on the opposite side of the homology, approximately equidistant from the center of the overlap. Fifty-seven individual G-418-resistant colonies representing the products of individual crosses were isolated, and the genomic DNAs containing the presumably integrated, functional recombinant neo genes were analyzed on Southern blots. A band consistent with the exchange of markers flanking the neo gene was present in 90% of the DNAs examined. In only one case was the pattern indicative of either a double crossover or a gene conversion event. These results support the idea that homologous extrachromosomal DNA fragments are joined through annealing of overlapping single-stranded ends. This DNA-joining phenomenon may represent the activity of cellular DNA repair enzymes; its relationship to genetic recombination occurring at the chromosomal level remains to be determined.     

Analysis of strand transfer and template switching mechanisms of DNA gap repair by homologous recombination in Escherichia coli: predominance of strand transfer

Daughter strand gaps formed upon interruption of replication at DNA lesions in Escherichiacoli can be repaired by either translesion DNA synthesis or homologous recombination (HR) repair. Using a plasmid-based assay system that enables discrimination between strand transfer and template switching (information copying) modes of HR gap repair, we found that approximately 80% of strand gaps were repaired by physical strand transfer from the donor, whereas approximately 20% appear to be repaired by template switching. HR gap repair operated on both small and bulky lesions and largely depended on RecA and RecF but not on the RecBCD nuclease. In addition, we found that HR was mildly reduced in cells lacking the RuvABC and RecG proteins involved in resolution of Holliday junctions. These results, obtained for the first time under conditions that detect the two HR gap repair mechanisms, provide in vivo high-resolution molecular evidence for the predominance of the strand transfer mechanism in HR gap repair. A small but significant portion of HR gap repair appears to occur via a template switching mechanism.     

Genetic exchange between homeologous sequences in mammalian chromosomes is averted by local homology requirements for initiation and resolution of recombination

We examined the mechanism by which recombination between imperfectly matched sequences (homeologous recombination) is suppressed in mammalian chromosomes. DNA substrates were constructed, each containing a thymidine kinase (tk) gene disrupted by insertion of an XhoI linker and referred to as a "recipient" gene. Each substrate also contained one of several "donor" tk sequences that could potentially correct the recipient gene via recombination. Each donor sequence either was perfectly homologous to the recipient gene or contained homeologous sequence sharing only 80% identity with the recipient gene. Mouse Ltk(-) fibroblasts were stably transfected with the various substrates and tk(+) segregants produced via intrachromosomal recombination were recovered. We observed exclusion of homeologous sequence from gene conversion tracts when homeologous sequence was positioned adjacent to homologous sequence in the donor but not when homeologous sequence was surrounded by homology in the donor. Our results support a model in which homeologous recombination in mammalian chromosomes is suppressed by a nondestructive dismantling of mismatched heteroduplex DNA (hDNA) intermediates. We suggest that mammalian cells do not dismantle mismatched hDNA by responding to mismatches in hDNA per se but rather rejection of mismatched hDNA appears to be driven by a requirement for localized homology for resolution of recombination.     

Intermolecular plasmid recombination in fibroblasts from humans with DNA damage-processing defects

We have evaluated the ability of immortalized human fibroblasts to recombine transfected plasmid DNA. A number of cell lines from normal individuals and from patients with DNA damage-processing defects were examined. Two plasmid recombination substrates were derived from pSV2neo and contained nonoverlapping deletions in the aminoglycoside phosphotransferase II gene. Intermolecular recombination was assessed by two methods after cotransfection. In a short-term, extrachromosomal recombination assay, low molecular weight DNA was extracted from the human cells 48 h after transfection, and recombinant plasmids were detected by transformation into appropriate indicator bacteria. In a long-term stable recombination assay the fibroblasts were cotransfected and G418-resistant colonies allowed to form. By the former assay all but two cultures were recombination-proficient, whereas all were recombination-proficient by the latter assay. The efficiency of transfection of human cells with plasmids appears to be a major variable affecting recombination. Recombination can be stimulated by uv irradiation of plasmid DNA prior to transfection. Cells from patients with Fanconi anemia, ataxia telangiectasia, and xeroderma pigmentosum complementation groups A, C, D, E, and G are not defective at intermolecular plasmid recombination.     

A Sister-Strand Exchange Mechanism for Reca-Independent Deletion of Repeated DNA Sequences in Escherichia Coli

In the genomes of many organisms, deletions arise between tandemly repeated DNA sequences of lengths ranging from several kilobases to only a few nucleotides. Using a plasmid-based assay for deletion of a 787-bp tandem repeat, we have found that a recA-independent mechanism contributes substantially to the deletion process of even this large region of homology. No Escherichia coli recombination gene tested, including recA, had greater than a fivefold effect on deletion rates. The recA-independence of deletion formation is also observed with constructions present on the chromosome. RecA promotes synapsis and transfer of homologous DNA strands in vitro and is indispensable for intermolecular recombination events in vivo measured after conjugation. Because deletion formation in E. coli shows little or no dependence on recA, it has been assumed that homologous recombination contributes little to the deletion process. However, we have found recA-independent deletion products suggestive of reciprocal crossovers when branch migration in the cell is inhibited by a ruvA mutation. We propose a model for recA-independent crossovers between replicating sister strands, which can also explain deletion or amplification of repeated sequences. We suggest that this process may be initiated as post-replicational DNA repair; subsequent strand misalignment at repeated sequences leads to genetic rearrangements.     

Restriction-Stimulated Homologous Recombination of Plasmids by the Rece Pathway of Escherichia Coli

To test the double-strand break (DSB) repair model in recombination by the RecE pathway of Escherichia coli, we constructed chimeric phages that allow restriction-mediated release of linear plasmid substrates of the bioluminescence recombination assay in infected EcoRI+ cells. Kinetics of DSB repair and expression of recombination products were followed by Southern hybridization and by the bioluminescence recombination assay, respectively. Plasmid recombinants were analyzed with restriction endonucleases. Our results indicate that a DSB can induce more than one type of RecE-mediated recombination. A DSB within the homology induced intermolecular recombination that followed the rules of the DSB repair model: (1) Recombination was enhanced by in vivo restriction. (2) Repair of the break depended on homologous sequences on the resident plasmid. (3) Break-repair was frequently associated with conversion of alleles that were cis to the break. (4) Conversion frequency decreased as the distance from the break increased. (5) Some clones contained a mixture of plasmid recombinants as expected by replication of a heteroduplex in the primary recombinant. The rules of the DSB repair model were not followed when recombination was induced by a DSB outside the homology. Both the cut and the uncut substrates were recipients in conversion events. Recombination events were associated with deletions that spanned the break site, but these deletions did not reach the homology. We propose that a break outside the homology may stimulate a RecE-mediated recombination pathway that does not involve direct participation of DNA ends in the homologous pairing reaction.     

The genomic instability associated with integrated simian virus 40 DNA is dependent on the origin of replication and early control region.

DNA rearrangements in the form of deletions and duplications are found within and near integrated simian virus 40 (SV40) DNA in nonpermissive cell lines. We have found that rearrangements also occur frequently with integrated pSV2neo plasmid DNA. pSV2neo contains the entire SV40 control region, including the origin of replication, both promoters, and the enhancer sequences. Linearized plasmid DNA was electroporated into X1, an SV40-transformed mouse cell line that expresses SV40 large T antigen (T Ag) and shows very frequent rearrangements at the SV40 locus, and into LMtk-, a spontaneously transformed mouse cell line that contains no SV40 DNA. Stability was analyzed by subcloning G-418-resistant clones and examining specific DNA fragments for alterations in size. Five independent X1 clones containing pSV2neo DNA were unstable at both the neo locus and the T Ag locus. By contrast, four X1 clones containing mutants of pSV2neo with small deletions in the SV40 core origin and three X1 clones containing a different neo plasmid lacking SV40 sequences were stable at the neo locus, although they were still unstable at the T Ag locus. Surprisingly, five independent LMtk- clones containing pSV2neo DNA were unstable at the neo locus. LMtk- clones containing origin deletion mutants were more stable but were not as stable as the X1 clones containing the same plasmid DNA. We conclude that the SV40 origin of replication and early control region are sufficient viral components for the genomic instability at sites of SV40 integration and that SV40 T Ag is not required.     

DNA sequence similarity requirements for interspecific recombination in Bacillus

Gene transfer in bacteria is notoriously promiscuous. Genetic material is known to be transferred between groups as distantly related as the Gram positives and Gram negatives. However, the frequency of homologous recombination decreases sharply with the level of relatedness between the donor and recipient. Several studies show that this sexual isolation is an exponential function of DNA sequence divergence between recombining substrates. The two major factors implicated in producing the recombinational barrier are the mismatch repair system and the requirement for a short region of sequence identity to initiate strand exchange. Here we demonstrate that sexual isolation in Bacillus transformation results almost exclusively from the need for regions of identity at both the 5' and 3' ends of the donor DNA strand. We show that, by providing the essential identity, we can effectively eliminate sexual isolation between highly divergent sequences. We also present evidence that the potential of a donor sequence to act as a recombinogenic, invasive end is determined by the stability (melting point) of the donor-recipient complex. These results explain the exponential relationship between sexual isolation and sequence divergence observed in bacteria. They also suggest a model for rapid spread of novel adaptations, such as antibiotic resistance genes, among related species.     

Initiation of DNA synthesis in the transfer origin region of RK2 by the plasmid-encoded primase: detection using defective M13 phage

The broad host range IncP (IncP1) plasmids of gram-negative bacteria encode DNA primases that are involved in conjugal DNA synthesis. The primase of RK2/RP4 is required for efficient DNA transfer to certain gram-negative bacteria, indicating that the enzyme primes complementary strand synthesis in the recipient. In vitro, the primase initiates synthesis of oligoribonucleotides at 3'-dGdT-5' dinucleotides on the template strand. In this report, replication-defective M13 phage are used to assay the ability of the RK2-encoded primase to initiate complementary strand synthesis in vivo on single-strand templates containing the RK2 origin of conjugal transfer (oriT) or the RK2 origin of vegetative replication (oriV). The results show that sequences from either strand of the oriT region serve as efficient substrates for the RK2 primase and can enhance the growth of the defective M13 vectors delta E101 and delta Elac to levels approaching wild-type. The primise-oriT interaction appeared specific, since neither the oriV sequence nor another RK2 region, trfB, significantly enhanced growth of the defective phage, either in the presence or in the absence of the primase. In contrast to ColEl and F, this study also shows that the oriV region of RK2 lacks sites that are recognized by the host-specified DNA priming systems. The results suggest that the oriT region contains sites on both DNA strands that are efficient substrates for the plasmid-encoded primase, facilitating initiation of complementary strand DNA synthesis in both donor and recipient during conjugation.     

Intron homing with limited exon homology. Illegitimate double-strand-break repair in intron acquisition by phage t4

The td intron of bacteriophage T4 encodes a DNA endonuclease that initiates intron homing to cognate intronless alleles by a double-strand-break (DSB) repair process. A genetic assay was developed to analyze the relationship between exon homology and homing efficiency. Because models predict exonucleolytic processing of the cleaved recipient leading to homologous strand invasion of the donor allele, the assay was performed in wild-type and exonuclease-deficient (rnh or dexA) phage. Efficient homing was supported by exon lengths of 50 bp or greater, whereas more limited exon lengths led to a precipitous decline in homing levels. However, extensive homology in one exon still supported elevated homing levels when the other exon was completely absent. Analysis of these "one-sided" events revealed recombination junctions at ectopic sites of microhomology and implicated nucleolytic degradation in illegitimate DSB repair in T4. Interestingly, homing efficiency with extremely limiting exon homology was greatly elevated in phage deficient in the 3'-5' exonuclease, DexA, suggesting that the length of 3' tails is a major determinant of the efficiency of DSB repair. Together, these results suggest that illegitimate DSB repair may provide a means by which introns can invade ectopic sites.     

Specific DNA cleavage mediated by the integrase of conjugative transposon Tn916.

The conjugative transposon Tn916 encodes a protein called INT(Tn916) which, based on DNA sequence comparisons, is a member of the integrase family of site-specific recombinases. Integrase proteins such as INT(lambda), FLP, and XERC/D that promote site-specific recombination use characteristic, conserved amino acid residues to catalyze the cleavage and ligation of DNA substrates during recombination. The reaction proceeds by a two-step transesterification reaction requiring the formation of a covalent protein-DNA intermediate. Different requirements for homology between recombining DNA sites during integrase-mediated site-specific recombination and Tn916 transposition suggest that INT(Tn916) may use a reaction mechanism different from that used by other integrase recombinases. We show that purified INT(Tn916) mediates specific cleavage of duplex DNA substrates containing the Tn916 transposon ends and adjacent bacterial sequences. Staggered cleavages occur at both ends of the transposon, resulting in 5' hydroxyl protruding ends containing coupling sequences. These are sequences that are transferred with the transposon from donor to recipient during conjugative transposition. The nature of the cleavage products suggests that a covalent protein-DNA linkage occurs via a residue of INT(Tn916) and the 3'-phosphate group of the DNA. INT(Tn916) alone is capable of executing the strand cleavage step required for recombination during Tn916 transposition, and this reaction probably occurs by a mechanism similar to that of other integrase family site-specific recombinases.     

Involvement of Rad52 in T‐DNA circle formation during Agrobacterium tumefaciens‐mediated transformation of Saccharomyces cerevisiae

Agrobacterium tumefaciens cells carrying a tumour inducing plasmid (Ti‐plasmid) can transfer a defined region of transfer DNA (T‐DNA) to plant cells as well as to yeast. This process of Agrobacterium‐mediated transformation (AMT) eventually results in the incorporation of the T‐DNA in the genomic DNA of the recipient cells. All available evidence indicates that T‐strand transfer closely resembles conjugal DNA transfer as found between Gram‐negative bacteria. However, where conjugal plasmid DNA transfer starts via relaxase‐mediated processing of a single origin of transfer (oriT), the T‐DNA is flanked by two imperfect direct border repeats which are both substrates for the Ti‐plasmid encoded relaxase VirD2. Yeast was used as a model system to investigate the requirements of the recipient cell for the formation of T‐DNA circles after AMT. It was found that, despite the absence of self‐homology on the T‐DNA, the homologous repair proteins Rad52 and Rad51 are involved in T‐DNA circle formation. A model is presented involving the formation of T‐DNA concatemers derived from T‐strands by a process of strand‐transfer catalysed by VirD2. These concatemers are then resolved into T‐DNA circles by homologous recombination in the recipient cell.     

Induction of direct repeat recombination by psoralen–DNA adducts in Saccharomyces cerevisiae: Defects in DNA repair increase gene copy number variation

Psoralen photoreaction produces covalent monoadducts and interstrand crosslinks in DNA. The interstrand DNA crosslinks are complex double strand lesions that require the involvement of multiple pathways for repair. Homologous recombination, which can carry out error-free repair, is a major pathway for crosslink repair; however, some recombination pathways can also produce DNA rearrangements. Psoralen photoreaction-induced recombination in yeast was measured using direct repeat substrates that can detect gene conversions, a form of conservative recombination, as well as deletions and triplications, which generate gene copy number changes. In repair-proficient cells the major products of recombination were gene conversions, along with substantial fractions of deletions. Deficiencies in DNA repair pathways increased non-conservative recombination products. Homologous recombination-deficient rad51, rad54, and rad57 strains had low levels of crosslink-induced recombination, and most products were deletions produced by single strand annealing. Nucleotide excision repair-deficient rad1 and rad2 yeast had increased levels of triplications, and rad1 cells had lower crosslink-induced recombination. Deficiencies in post-replication repair increased crosslink-induced recombination and gene copy number changes. Loss of REV3 function, in the error-prone branch, and of RAD5 and UBC13, in the error-free branch, produced moderate increases in deletions and triplications; rad18 cells, deficient in both post-replication repair sub-pathways, exhibited hyperrecombination, with primarily non-conservative products. Proper functioning of all the DNA repair pathways tested was required to maintain genomic stability and avoid gene copy number variation in response to interstrand crosslinks.     

Effect of double-strand breaks on homologous recombination in mammalian cells and extracts.

We examined the effect of double-strand breaks on homologous recombination between two plasmids in human cells and in nuclear extracts prepared from human and rodent cells. Two pSV2neo plasmids containing nonreverting, nonoverlapping deletions were cotransfected into cells or incubated with cell extracts. Generation of intact neo genes was monitored by the ability of the DNA to confer G418r to cells or Neor to bacteria. We show that double-strand breaks at the sites of the deletions enhanced recombination frequency, whereas breaks outside the neo gene had no effect. Examination of the plasmids obtained from experiments involving the cell extracts revealed that gene conversion events play an important role in the generation of plasmids containing intact neo genes. Studies with plasmids carrying multiple polymorphic genetic markers revealed that markers located within 1,000 base pairs could be readily coconverted. The frequency of coconversion decreased with increasing distance between the markers. The plasmids we constructed along with the in vitro system should permit a detailed analysis of homologous recombinational events mediated by mammalian enzymes.     

Gene targeting with a replication-defective adenovirus vector.

Wide application of the gene-targeting technique has been hampered by its low level of efficiency. A replication-defective adenovirus vector was used for efficient delivery of donor DNA in order to bypass this problem. Homologous recombination was selected between a donor neo gene inserted in the adenovirus vector and a target mutant neo gene on a nuclear papillomavirus plasmid. These recombinant adenoviruses allowed gene transfer to 100% of the treated cells without impairing their viability. Homologous recombinants were obtained at a level of frequency much higher than that obtained by electroporation or a calcium phosphate procedure. The structure of the recombinants was analyzed in detail after recovery in an Escherichia coli strain. All of the recombinants examined had experienced a precise correction of the mutant neo gene. Some of them had a nonhomologous rearrangement of their sequences as well. One type of nonhomologous recombination took place at the end of the donor-target homology. The vector adenovirus DNA was inserted into some of the products obtained at a high multiplicity of infection. The insertion was at the end of the donor-target homology with a concomitant insertion of a 10-bp-long filler sequence in one of the recombinants. The possible relationship between these rearrangements and the homologous recombination is discussed. These results demonstrate the applicability of adenovirus-mediated gene delivery in gene targeting and gene therapy.