Genetic and Physical Analysis of Plasmid Recombination in Recb Recc Sbcb and Recb Recc Sbca Escherichia Coli K-12 Mutants

The effect of mutations in known recombination genes (recA, recB, recC, recE, recF, recJ, recN, recO, recQ and ruv) on intramolecular recombination of plasmids was studied in recB recC sbcB and recB recC sbcA Escherichia coli mutants. The rate of recombination of circular dimer plasmids was at least 1000-fold higher in recB recC sbcB or recB recC sbcA mutants as compared to wild-type cells. The rate was decreased by mutations in recA, recF, recJ, recO, ruv or mutS in recB recC sbcB mutants, and by mutations in recE, recN, recO, recQ, ruv or mutS in recB recC sbcA mutants. In addition to measuring the recombination rate of circular dimer plasmids, the recombination-mediated transformation of linear dimer plasmids was also studied. Linear dimer plasmids transformed recB recC sbcB and recB recC sbcA mutants 20- to 40-fold more efficiently than wild-type cells. The transformation efficiency of linear dimer plasmids in recB recC sbcB mutants was decreased by mutations in recA, recF, recJ, recO, recQ or lexA (lexA3). In recB recC sbcA mutants the transformation efficiency of linear dimers was decreased only by a recE mutation. Physical analysis of linear dimer- or circular dimer-transformed recB recC sbcB mutants revealed that all transformants contained recombinant monomer genotypes. This suggests that recombination in recB recC sbcB cells is very efficient.     

Heteroduplex Strand-Specificity in Restriction-Stimulated Recombination by the Rece Pathway of Escherichia Coli

The RecE recombination pathway is active in Escherichia coli recB recC sbcA mutants. To isolate and characterize products and intermediates of RecE-mediated, break-induced, intramolecular recombination, we infected recB recC sbcA mutants, expressing EcoRI endonuclease, with chimeric λ phages that allow EcoRI-mediated release of cloned linear recombination substrates. Substrates with direct terminal repeats recombined to yield a circular product with one copy of the repeated sequence. Some recombinants were heteroallelic for the recombining markers. Markers distant to the break were recovered in the circular product at a higher frequency than markers close to the break. To examine the heteroduplex structures that may have yielded the heteroallelic recombinants, nonreplicative substrates were employed. Some of the nonreplicative recombination products contained heteroduplexes, with a strong bias for paired strands ending 3' at the break. This strand bias in heteroduplex formation is consistent with recombination models that postulate homologous pairing of protruding 3' single-stranded ends.     

Genetic analysis of double-strand break repair in Escherichia coli.

We had reported that a double-strand gap (ca. 300 bp long) in a duplex DNA is repaired through gene conversion copying a homologous duplex in a recB21 recC22 sbcA23 strain of Escherichia coli, as predicted on the basis of the double-strand break repair models. We have now examined various mutants for this repair capacity. (i) The recE159 mutation abolishes the reaction in the recB21C22 sbcA23 background. This result is consistent with the hypothesis that exonuclease VIII exposes a 3'-ended single strand from a double-strand break. (ii) Two recA alleles, including a complete deletion, fail to block the repair in this recBC sbcA background. (iii) Mutations in two more SOS-inducible genes, recN and recQ, do not decrease the repair. In addition, a lexA (Ind-) mutation, which blocks SOS induction, does not block the reaction. (iv) The recJ, recF, recO, and recR gene functions are nonessential in this background. (v) The RecBCD enzyme does not abolish the gap repair. We then examined genetic backgrounds other than recBC sbcA, in which the RecE pathway is not active. We failed to detect the double-strand gap repair in a rec+, a recA1, or a recB21 C22 strain, nor did we find the gap repair activity in a recD mutant or in a recB21 C22 sbcB15 sbcC201 mutant. We also failed to detect conservative repair of a simple double-strand break, which was made by restriction cleavage of an inserted linker oligonucleotide, in these backgrounds. We conclude that the RecBCD, RecBCD-, and RecF pathways cannot promote conservative double-strand break repair as the RecE and lambda Red pathways can.     

Repair of ultraviolet light-damaged deoxyribonucleic acid in sbc-A strains of Escherichia coli K-12.

An Escherichia coli strain carrying both rec+ and sbcA has been constructed. Repair of ultraviolet light-induced deoxyribonucleic acid damage was examined by measuring survival and thymine-dimer excision in the rec+ sbcA strain as well as rec+ sbcA+ and recB recC sbcA strains. The sbcA mutation restores normal survival in both recB recC uvrB and recB recC uvr+ strains. Excision of thymine-containing dimers does not occur in uvrB mutants, regardless of the rec or sbcA genotype. Survival, after ultraviolet-light damage, of a rec+ sbcA strain is quantitatively similar to rec+ sbcA+ and recB recC sbcA strains.     

Genetic analysis of the recG locus of Escherichia coli K-12 and of its role in recombination and DNA repair.

We describe a transposon insertion that reduces the efficiency of homologous recombination and DNA repair in Escherichia coli. The insertion, rec-258, was located between pyrE and dgo at min 82.1 on the current linkage map. On the basis of linkage to pyrE and complementation studies with the cloned rec+ gene, rec-258 was identified as an allele of the recG locus first reported by Storm et al. (P. K. Storm, W. P. M. Hoekstra, P. G. De Haan, and C. Verhoef, Mutat. Res. 13:9-17, 1971). The recG258 mutation confers sensitivity to mitomycin C and UV light and a 3- to 10-fold deficiency in conjugational recombination in wild-type, recB recC sbcA, and recB recC sbcB sbcC genetic backgrounds. It does not appear to affect plasmid recombination in the wild-type. A recG258 single mutant is also sensitive to ionizing radiation. The SOS response is induced normally, although the basal level of expression is elevated two- to threefold. Further genetic studies revealed that recB recG and recG recJ double mutants are much more sensitive to UV light than the respective single mutants in each case. However, no synergistic interactions were discovered between recG258 and mutations in recF, recN, or recQ. It is concluded that recG does not fall within any of the accepted groups of genes that affect recombination and DNA repair.     

Genetic recombination of bacterial plasmid DNA. Physical and genetic analysis of the products of plasmid recombination in Escherichia coli

Derivatives of plasmid pBR322 DNA containing tet mutations were constructed by inserting XhoI linkers at various sites in the tetracycline resistance gene. Monomer plasmids containing either the tet-10 allele located at nucleotide position 23 or the tet-14 allele located at nucleotide position 1267 were used to construct a circular dimer containing one copy of each allele and a circular trimer containing one copy of the tet-10 allele and two copies of the tet-14 allele. Genetic recombination of these plasmid DNAs to produce a functional tetracycline resistance gene could be detected as the production of tetracycline-resistant progeny during the growth of transformants or using a restriction mapping assay which detected the rearrangement of the mutant alleles. The structure of individual tetracycline-resistant recombination products was determined by restriction mapping. This analysis suggested that as many as 70% of the plasmid recombination events in Escherichia coli AB1157 could have involved gene conversion events. The formation of these recombination products was most easily predicted by a model involving figure 8 recombination intermediates and the formation of symmetric regions of heteroduplex. Recombination in JC10287 delta(srlR-recA)304 occurred at 5% of the wild-type frequency and appeared to occur by a similar mechanism. Recombination in JC9604 recA56 recB21 recC22 sbcA23 occurred at 20 times the wild-type frequency and appeared to involve multiple independent recombination events.     

Repair of cross-linked DNA and survival of Escherichia coli treated with psoralen and light: effects of mutations influencing genetic recombination and DNA metabolism.

Repair of cross-linked DNA was studied in Escherichia coli strains carrying mutations affecting DNA metabolism. In wild-type cells, DNA strands cut during cross-link removal were rejoined during a subsequent incubation into high-molecular-weight molecules. This rejoining was dependent on gene products involved in genetic recombination. A close correlation was found relating recombination proficiency, the rate of strand rejoining, and formation of viable progeny after DNA cross-linking by treatment with psoralen and light. Wild-type cells and other mutants which were Rec+ (sbcB, recL, recL sbcB, recB recC sbcA, recB recC sbcB, xthA1, and xthA11) rejoined cut DNA strands at a rate of 0.8 +/- 0.1 min -1 at 37 degrees C and survived 53 to 71 cross-links per chromosome. recB, recC, recB recC, recF, or polA strains showed reduced rates of strand rejoining and survived 4 to 13 cross-links per chromosome. Recombination-deficient strains (recA, recB recC sbcB recF, recB recL) and lexA failed to rejoin DNA strands after crosslink removal and were unable to form colonies after treatments producing as few as one to two cross-links per chromosome. Strand rejoining occurred normally in cells with mutations affecting DNA replication (dnaA, danB, dnaG, and dnaE) under both permissive and nonpermissive conditions for chromosome replication. In a polA polB dnaE strain strand rejoining occurred at 32 degree C but not at 42 degree C, indicating that some DNA synthesis was required for formation of intact recombinant molecules.     

Differential Thermolability of Exonuclease and Endonuclease Activities of the recBC Nuclease Isolated from Thermosensitive recB and recC Mutants

The recBC nuclease (also called exonuclease V) has been partially purified from Escherichia coli K-12 strains carrying the thermosensitive recB270, recC271, and recB270 recC271 mutations. Of the multiple activities associated with the enzyme, only the adenosine 5'-triphosphate-dependent exonucleolytic hydrolysis of duplex deoxyribonucleic acid (DNA) is abnormally thermolabile. The exo- and endonucleolytic degradation of single-stranded DNA is no more thermosensitive than that catalyzed by the wild-type enzyme. These results suggest that the defects in genetic recombination, DNA repair, and the maintenance of cell viability observed in recBC mutants in vivo result primarily from the specific loss of adenosine 5'-triphosphate-dependent exonuclease active on duplex DNA.     

Synapsis-Mediated Fusion of Free DNA Ends Forms Inverted Dimer Plasmids in Yeast

When yeast (Saccharomyces cerevisiae) is transformed with linearized plasmid DNA and the ends of the plasmid do not share homology with the yeast genome, circular inverted (head-to-head) dimer plasmids are the principal product of repair. By measurements of the DNA concentration dependence of transformation with a linearized plasmid, and by transformation with mixtures of genetically marked plasmids, we show that two plasmid molecules are required to form an inverted dimer plasmid. Several observations suggest that homologous pairing accounts for the head-to-head joining of the two plasmid molecules. First, an enhanced frequency of homologous recombination is detected when genetically marked plasmids undergo end-to-end fusion. Second, when a plasmid is linearized within an inverted repeat, such that its ends could undergo head-to-tail homologous pairing, it is repaired by intramolecular head-to-tail joining. Last, in the joining of homologous linearized plasmids of different length, a shorter molecule can acquire a longer plasmid end by homologous recombination. The formation of inverted dimer plasmids may be related to some forms of chromosomal rearrangement. These might include the fusion of broken sister chromatids in the bridge-breakage-fusion cycle and the head-to-head duplication of genomic DNA at the sites of gene amplifications.     

Genetic recombination: recA protein promotes homologous pairing between duplex DNA molecules without strand unwinding.

The recA protein of Escherichia coli promotes pairing in vitro between covalent circular duplex DNA and homologous circular duplex DNA containing a single stranded region. We have used a filter binding assay to investigate the frequency of homologous pairing between gapped and intact duplex DNA when unwinding of the free 3' and 5' ends of the gapped molecules was blocked. In order to obtain DNA without free ends, the gapped DNA was treated with trimethylpsoralen and 360 nm light so as to introduce about 6 crosslinks per DNA molecule and the double stranded regions on either side of the gaps were then digested up to the first crosslinks with exonuclease III and lambda exonuclease. This treatment did not diminish the frequency of homologous pairing, an observation which is difficult to reconcile with models for recombination requiring strand unwinding before pairing.     

Evidence that recBC-dependent degradation of duplex DNA in Escherichia coli recD mutants involves DNA unwinding.

Infection of Escherichia coli with phage T4 gene 2am was used to transport 3H-labeled linear duplex DNA into cells to follow its degradation in relation to the cellular genotype. In wild-type cells, 49% of the DNA was made acid soluble within 60 min; in recB or recC cells, only about 5% of the DNA was made acid soluble. Remarkably, in recD cells about 25% of the DNA was rendered acid soluble. The DNA degradation in recD cells depended on intact recB and recC genes. The degradation in recD cells was largely decreased by mutations in recJ (which eliminates the 5' single-strand-specific exonuclease coded by this gene) or xonA (which abolishes the 3' single-strand-specific exonuclease I). In a recD recJ xonA triple mutant, the degradation of linear duplex DNA was roughly at the level of a recB mutant. Results similar to those with the set of recD strains were also obtained with a recC++ mutant (in which the RecD protein is intact but does not function) and its recJ, xonA, and recJ xonA derivatives. The observations provide evidence for a recBC-dependent DNA-unwinding activity that renders unwound DNA susceptible to exonucleolytic degradation. It is proposed that the DNA-unwinding activity causes the efficient recombination, DNA repair, and SOS induction (after application of nalidixic acid) in recD mutants. The RecBC helicase indirectly detected here may have a central function in Chi-dependent recombination and in the recombinational repair of double-strand breaks by the RecBCD pathway.     

Mutation-dependent suppression of recB21 recC22 by a region cloned from the Rac prophage of Escherichia coli K-12.

Using pBR322 as a vector, we cloned a 5.95-kilobase fragment of the Rac prophage together with 1.70 kilobases of a flanking Escherichia coli chromosome sequence. The resulting plasmid (pRAC1) was unable to suppress the mitomycin and UV sensitivity and recombination deficiency of a recB21 recC22 strain. Five spontaneous mitomycin-resistant derivatives contained deletion mutant plasmids. These plasmids also suppressed the UV sensitivity and recombination deficiency of their recB21 recC22 hosts. All five deletions were contained within a 2.45-kilobase EcoRI-to-HindIII segment of the plasmid. By substituting the corresponding 2.45-kilobase EcoRI-toHindIII fragments of Rac prophage isolated from sbcA+, sbcA6, and sbcA23 strains for the shortened segment of one of the deletion mutant plasmids, we were able to show that sbcA mutations map in this region. Also in this region is the site (or closely linked sites) at which previous studies had shown that insertion of Tn5 and IS50 leads to suppression of recB21 recC22. The sequence in this region that must be altered or circumvented to allow suppression is discussed. Also presented are data correlating the expression of nuclease activity with the degree of suppression.     

Identification and characterization of the Escherichia coli RecT protein, a protein encoded by the recE region that promotes renaturation of homologous single-stranded DNA.

Recombination of plasmid DNAs and recombination of bacteriophage lambda red mutants in recB recC sbcA Escherichia coli mutants, in which the recE region is expressed, do not require recA. The recE gene is known to encode exonuclease VIII (exoVIII), which is an ATP-independent exonuclease involved in the RecE pathway of recombination. A 33,000-molecular-weight (MW) protein was observed to be coexpressed with both exoVIII and a truncated version of exoVIII, pRac3 exo, when they were overproduced under the control of strong promoters. We have purified this 33,000-MW protein (p33) and demonstrated by protein sequence analysis that it is encoded by the same coding sequence that encodes the C-terminal 33,000-MW portion of exoVIII. p33 is expressed independently of exoVIII but is probably translated from the same mRNA. p33 was found to bind to single-stranded DNA and also to promote the renaturation of complementary single-stranded DNA. It appears that p33 is functionally analogous to the bacteriophage lambda beta protein, which may explain why RecE pathway recombination does not require recA.     

Transformation of yeast with linearized plasmid DNA. Formation of inverted dimers and recombinant plasmid products

The molecular products of DNA double strand break repair were investigated after transformation of yeast (Saccharomyces cerevisiae) with linearized plasmid DNA. DNA of an autonomous yeast plasmid cleaved to generate free ends lacking homology with the yeast genome, when used in transformation along with sonicated non-homologous carrier DNA, gave rise to transformants with high frequency. Most of these transformants were found to harbor a head-to-head (inverted) dimer of the linearized plasmid. This outcome of transformation contrasts with that observed when the carrier DNA is not present. Transformants occur at a much reduced frequency and harbor either the parent plasmid or a plasmid with deletion at the site of the cleavage. When the linearized plasmid is introduced along with sonicated carrier DNA and a homologous DNA restriction fragment that spans the site of plasmid cleavage, homologous recombination restores the plasmid to its original circular form. Inverted dimer plasmids are not detected. This relationship between homologous recombination and a novel DNA transaction that yields rearrangement could be important to the cell, as the latter could lead to a loss of gene function and lethality.     

Recombination of bacteriophage lambda in recD mutants of Escherichia coli

RecBCD enzyme is centrally important in homologous recombination in Escherichia coli and is the source of ExoV activity. Null alleles of either the recB or the recC genes, which encode the B and C subunits, respectively, manifest no recombination and none of the nuclease functions characteristic of the holoenzyme. Loss of the D subunit, by a recD mutation, likewise results in loss of ExoV activity. However, mutants lacking the D subunit are competent for homologous recombination. We report that the distribution of exchanges along the chromosome of Red-Gam-phage lambda is strikingly altered by recD null mutations in the host. When lambda DNA replication is blocked, recombination in recD mutant strains is high near lambda's right end. In contrast, recombination in isogenic recD+ strains is approximately uniform along lambda unless the lambda chromosome contains a chi sequence. Recombination in recD mutant strains is focused toward the site of action of a type II restriction enzyme acting in vivo on lambda. The distribution of exchanges in isogenic recD+ strains is scarcely altered by the restriction enzyme (unless the phage contains an otherwise silent chi). The distribution of exchanges in recD mutants is strongly affected by lambda DNA replication. The distribution of exchanges on lambda growing in rec+ cells is not influenced by DNA replication. The exchange distribution along lambda in recD mutant cells is independent of chi in a variety of conditions. Recombination in rec+ cells is chi influenced. Recombination in recD mutants depends on recC function, occurs in strains deleted for rac prophage, and is independent of recJ, which is known to be required for lambda recombination via the RecF pathway. We entertain two models for recombination in recD mutants: (i) recombination in recD mutants may proceed via double-chain break--repair, as it does in lambda's Red pathway and E. coli's RecE pathway; (ii) the RecBC enzyme, missing its D subunit, is equivalent to the wild-type, RecBCD, enzyme after that enzyme has been activated by a chi sequence.     

Double-Stranded Gap Repair of DNA by Gene Conversion in Escherichia Coli

We demonstrated repair of a double-stranded DNA gap through gene conversion by a homologous DNA sequence in Escherichia coli. We made a double-stranded gap in one of the two regions of homology in an inverted orientation on a plasmid DNA molecule and introduced it into an E. coli strain which has the RecE system of recombination (genotype; sbcA23 recB21 recC22). We detected repair products by genetic selection. The repair products were those expected by the double-strand-gap repair model. Gene conversion was frequently accompanied by crossing over of the flanking sequences as in eukaryotes. This double-strand gap repair mechanism can explain plasmid recombination in the absence of an artificial double-stranded break reported in a companion study by Yamamoto et al.     

Kinetics of homologous pairing promoted by RecA protein: effects of ends and internal sites in DNA

When recA protein was preincubated with single-stranded DNA in the presence of an ATP-regenerating system prior to the addition of homologous duplex DNA, a slow presynaptic step was eliminated, and the subsequent homologous pairing was revealed as a reaction whose rate exceeds by 1 or 2 orders of magnitude the calculated rate of spontaneous renaturation in 0.15 M NaCl at Tm -25 degrees C. The pairing reaction displayed saturation kinetics with respect to both single-stranded and double-stranded DNA, indicating the existence of a rate-limiting enzyme-substrate complex. The signal observed in the assay of the pairing reaction was due to pairing at free homologous ends of the duplex DNA, as well as pairing in the middle of the duplex molecule, away from a free end. The apparent rate of pairing of circular single strands with linear duplex DNA was equal to the sum of the rates of pairing at sites located at either end of the duplex DNA or at interior sites, but the apparent rates attributable to ends were greater, and nicks also stimulated the apparent rate.     

Chromosome Transfer from F-lac+ Strains of Escherichia coli K-12 Mutant at recA, recB, or recC

The frequency of chromosome transfer from various recombination-deficient F-lac(+) donor strains was estimated by standardizing the yield of conjugants receiving a male chromosomal marker against the level of episome transfer in the mating mixture. The efficiency of chromosome transfer from newly formed F-lac(+) cells carrying recB21 or recC22 was more than 50% of the wild-type value, although it was about 10 and 20%, respectively, if the male cell lines had become established. In contrast, recA13 donors transmitted the chromosome with less than 10(-4) of the normal frequency. If chromosome transfer from F-lac(+) strains reflects the cutting and subsequent joining of homologous single strands of episomal and chromosomal deoxyribonucleic acid by recombination, these results imply that the completed unions are not made in recA cells, but can be effected with more than 50% of normal efficiency in newly formed partial diploids mutant at either recB or recC. Thus, the defective stage in recA mutants may precede strand joining, whereas the deficiency in recB or recC cells may involve a later step in recombinant formation.     

Gene conversion in Escherichia coli: the recF pathway for resolution of heteroduplex DNA.

The independent repair of mismatched nucleotides present in heteroduplex DNA has been used to explain gene conversion and map expansion after general genetic recombination. We have constructed and purified heteroduplex plasmid DNAs that contain heteroallelic 10-base-pair insertion-deletion mismatches. These DNA substrates are similar in structure to the heteroduplex DNA intermediates that have been proposed to be produced during the genetic recombination of plasmids. These DNA substrates were transformed into wild-type and mutant Escherichia coli strains, and the fate of the heteroduplex DNA was determined by both restriction mapping and genetic tests. Independent repair events that yielded a wild-type Tetr gene were observed at a frequency of approximately 1% in both wild-type and recB recC sbcB mutant E. coli strains. The independent repair of small insertion-deletion-type mismatches separated by 1,243 base pairs was found to be reduced by recF, recJ, and ssb single mutations in an otherwise wild-type genetic background and reduced by recF, recJ, and recO mutations in a recB recC sbcB genetic background (the ssb mutation was not tested in the latter background). Independent repair of small insertion-deletion-type mismatched nucleotides that were as close as 312 nucleotides apart was observed. There was no apparent bias in favor of the insertion or deletion of mutant sequences.     

Characteristics of Some Multiply Recombination-Deficient Strains of Escherichia coli

Strains of Escherichia coli have been made carrying lesions in more than one gene determining recombination. The following genotypes were constructed and verified: recC22 recB21 recA(+), recC22 recB21 recA13, recC22 recB(+)recA13, and recC(+)recB21 recA13. All multiple rec(-) strains carrying recA13 were similar to AB2463, which carries recA13 alone, in their UV sensitivities, recombination deficiencies, and inabilities to induce lambda phage in a lysogen. However, whereas AB2463 shows a high rate of ultraviolet (UV)-induced deoxyribonucleic acid (DNA) breakdown, the multiple rec(-) strains showed the low level characteristic of strains carrying recC22 or recB21 alone. The strain carrying both recC22 and recB21 was similar in all properties to the single mutants, suggesting that both gene products act in the same part of the recombination and UV repair pathways. It is concluded that in a Rec(+) strain, the recA(+) product acts to inhibit DNA breakdown determined by the recC(+) and recB(+) products.