Homologous pairing in duplex DNA regions and the formation of four-stranded paranemic joints promoted by RecA protein. Effects of gap length and negative superhelicity

RecA protein catalyzes homologous pairing of partially single-stranded duplex DNA and fully duplex DNA to form stable joint molecules. We constructed circular duplex DNA with various defined gap lengths and studied the pairing reaction between the gapped substrate with fully double-stranded DNA. The reaction required a stoichiometric amount of RecA protein, and the optimal reaction was achieved at a ratio of 1 RecA monomer per 4 base pairs. The length of the gap, ranging from 141 to 1158 nucleotides, had little effect on the efficiency of homologous pairing. By using a circular gapped duplex DNA prepared from the chimeric phage M13Gori1, we were able to show the formation of nonintertwined or paranemic joints in duplex regions between the gapped and fully duplex molecules. The formation of such paranemic joints occurred efficiently and included nearly all of the DNA in the reaction mixture. The reaction required negative superhelicity, and pairing was greatly reduced with linear or nicked circular DNA. We conclude that one functional role of the single-stranded gap is for facilitating the binding of RecA protein to the duplex region of the gapped DNA. Once the nucleoprotein filament is formed, homologous pairing between the gapped and fully duplex DNA can take place anywhere along the length of the nucleoprotein complex.     

Absorption and fluorescence studies of the binding of the recA gene product from E. coli to single-stranded and double-stranded DNA. Ionic strength dependence

The binding of the recA gene product from E. coli to double-stranded and single-stranded nucleic acids has been investigated by following the change in melting temperature of duplex DNA and the fluorescence of single-stranded DNA or poly(dA) modified by reaction with chloroacetaldehyde. At low ionic strength, in the absence of Mg2+ ions, RecA protein binds preferentially to duplex DNA or poly(dA-dT). This leads to an increase of the DNA melting temperature. Stabilization of duplex DNA decreases when ionic strength or pH increases. In the presence of Mg2+ ions, preferential binding to single-stranded polynucleotides is observed. Precipitation occurs when duplex DNA begins to melt in the presence of RecA protein. From competition experiments, different single-stranded and double-stranded polydeoxynucleotides can be ranked according to their ability to bind RecA protein. Structural changes induced in nucleic acids upon RecA binding are discussed together with conformational changes induced in RecA protein upon magnesium binding.     

RecA protein-promoted homologous pairing between duplex molecules: functional role of duplex regions of gapped duplex DNA

RecA protein promotes homologous pairing and symmetrical strand exchange between partially single-stranded duplex DNA and fully duplex molecules. We constructed circular gapped DNA with a defined gap length and studied the pairing reaction between the gapped substrate and fully duplex DNA. RecA protein polymerizes onto the single-stranded and duplex regions of the gapped DNA to form a nucleoprotein filament. The formation of such filaments requires a stoichiometric amount of RecA protein. Both the rate and yield of joint molecule formation were reduced when the pairing reaction was carried out in the presence of a sub-saturating amount of RecA protein. The amount of RecA protein required for optimal pairing corresponds to the binding site size of RecA protein at saturation on duplex DNA. The result suggests that in the 4-stranded system the single-stranded as well as the duplex regions are involved in pairing. By using fully duplex DNA that shares different lengths and regions of homology with the gapped molecule, we directly showed that the duplex region of the gapped DNA increased both the rate and yield of joint molecule formation. The present study indicates that even though strand exchange in the 4-stranded system must require the presence of a single-stranded region, the pairing that occurs in duplex regions between DNA molecules is functionally significant and contributes to the overall activity of the gapped DNA.     

RNA-DNA hybridization promoted by E. coli RecA protein.

RecA protein of E. coli plays a central regulatory role that is induced by damage to DNA and results in the inactivation of LexA repressor. In vitro, RecA protein binds preferentially to single-stranded DNA to form a nucleoprotein filament that can recognize homology in naked duplex DNA and promote extensive strand exchange. Although RecA protein shows little tendency at neutral pH to bind to RNA, we found that it nonetheless catalyzed at 37 degrees C the hybridization of complementary RNA and single-stranded DNA sequences. Hybrids made by RecA protein at 37 degrees C appeared indistinguishable from ones prepared by thermal annealing. RNA-DNA hybridization by RecA protein at neutral pH required, as does RecA-promoted homologous pairing, optimal conditions for the formation of RecA nucleoprotein filaments. The cosedimentation of RNA with those filaments further paralleled observations made on the formation of networks of nucleoprotein filaments with double-stranded DNA, an instrumental intermediate in homologous pairing in vitro. These similarities with the pairing reaction support the view that RecA protein acts specifically in the hybridization reaction.     

Analysis of double stranded DNA-dependent activities of Deinococcus radiodurans RecA protein

In this study, the double-stranded DNA-dependent activities of Deinococcus radiodurans RecA protein (Dr RecA) were characterized. The interactions of the Dr RecA protein with double-stranded DNA were determined, especially dsDNA-dependent ATP hydrolysis by the Dr RecA protein and the DNA strand exchange reaction, in which multiple branch points exist on a single RecA protein-DNA complex. A nucleotide cofactor (ATP or dATP ) was required for the Dr RecA protein binding to duplex DNA. In the presence of dATP, the nucleation step in the binding process occurred more rapidly than in the presence of ATP. Salts inhibited the binding of the Dr RecA protein to double-stranded DNA. Double-stranded DNA-dependent ATPase activities showed a different sensitivity to anion species. Glutamate had only a minimal effect on the double-stranded DNA-dependent ATPase activities, up to a concentration of 0.7 M. In the competition experiment for Dr RecA protein binding, the Dr RecA protein manifested a higher affinity to double-stranded DNA than was observed for single-stranded DNA.     

C-terminal deletions of the Escherichia coli RecA protein. Characterization of in vivo and in vitro effects

A set of C-terminal deletion mutants of the RecA protein of Escherichia coli, progressively removing 6, 13, 17, and 25 amino acid residues, has been generated, expressed, and purified. In vivo, the deletion of 13 to 17 C-terminal residues results in increased sensitivity to mitomycin C. In vitro, the deletions enhance binding to duplex DNA as previously observed. We demonstrate that much of this enhancement involves the deletion of residues between positions 339 and 346. In addition, the C-terminal deletions cause a substantial upward shift in the pH-reaction profile of DNA strand exchange reactions. The C-terminal deletions of more than 13 amino acid residues result in strong inhibition of DNA strand exchange below pH 7, where the wild-type protein promotes a proficient reaction. However, at the same time, the deletion of 13-17 C-terminal residues eliminates the reduction in DNA strand exchange seen with the wild-type protein at pH values between 7.5 and 9. The results suggest the existence of extensive interactions, possibly involving multiple salt bridges, between the C terminus and other parts of the protein. These interactions affect the pK(a) of key groups involved in DNA strand exchange as well as the direct binding of RecA protein to duplex DNA.     

RecA protein of Mycobacterium tuberculosis possesses pH-dependent homologous DNA pairing and strand exchange activities: implications for allele exchange in mycobacteria

To gain insights into inefficient allele exchange in mycobacteria, we compared homologous pairing and strand exchange reactions promoted by RecA protein of Mycobacterium tuberculosis to those of Escherichia coli RecA protein. The extent of single-stranded binding protein (SSB)-stimulated formation of joint molecules by MtRecA was similar to that of EcRecA over a wide range of pH values. In contrast, strand exchange promoted by MtRecA was inhibited around neutral pH due to the formation of DNA networks. At higher pH, MtRecA was able to overcome this constraint and, consequently, displayed optimal strand exchange activity. Order of addition experiments suggested that SSB, when added after MtRecA, was vital for strand exchange. Significantly, with shorter duplex DNA, MtRecA promoted efficient strand exchange without network formation in a pH-independent fashion. Increase in the length of duplex DNA led to incomplete strand exchange with concomitant rise in the formation of intermediates and networks in a pH-dependent manner. Treatment of purified networks with S1 nuclease liberated linear duplex DNA and products, consistent with a model in which the networks are formed by the invasion of hybrid DNA by the displaced linear single-stranded DNA. Titration of strand exchange reactions with ATP or salt distinguished a condition under which the formation of networks was blocked, but strand exchange was not significantly affected. We discuss how these results relate to inefficient allele exchange in mycobacteria.     

By searching processively RecA protein pairs DNA molecules that share a limited stretch of homology

RecA protein promotes homologous pairing by a reaction in which the protein first binds stoichiometrically to single-stranded DNA in a slow presyn-aptic step, and then conjoins single-stranded and duplex DNA, thereby forming a ternary complex. RecA protein did not pair molecules that shared only 30 bp homology, but, with full efficiency, it paired circular single-stranded and linear duplex molecules in which homology was limited to 151 bp at one end of the duplex DNA. The initial rate of the pairing reaction was directly related to the length of the heterologous part of the duplex DNA, which we varied from 0 to 3060 base pairs. Since interactions involving the heterologous part of a molecule speed the location of a small homologous region, we conclude that RecA protein promotes homologous alignment by a processive mechanism involving relative motion of conjoined molecules within the ternary complex.     

RecA protein-promoted homologous pairing and strand exchange between intact and partially single-stranded duplex DNA.

In the pairing reaction between circular gapped and fully duplex DNA, RecA protein first polymerizes on the gapped DNA to form a nucleoprotein filament. Conditions that removed the formation of secondary structure in the gapped DNA, such as addition of Escherichia coli single-stranded DNA binding protein or preincubation in 1 mM-MgCl2, optimized the binding of RecA protein and increased the formation of joint molecules. The gapped duplex formed stable joints with fully duplex DNA that had a 5' or 3' terminus complementary to the single-stranded region of the gapped molecule. However, the joints formed had distinct properties and structures depending on whether the complementary terminus was at the 5' or 3' end. Pairing between gapped DNA and fully duplex linear DNA with a 3' complementary terminus resulted in strand displacement, symmetric strand exchange and formation of complete strand exchange products. By contrast, pairing between gapped and fully duplex DNA with a 5' complementary terminus produced a joint that was restricted to the gapped region; there was no strand displacement or symmetric strand exchange. The joint formed in the latter reaction was likely a three-stranded intermediate rather than a heteroduplex with the classical Watson-Crick structure. We conclude that, as in the three-strand reaction, the process of strand exchange in the four-strand reaction is polar and progresses in a 5' to 3' direction with respect to the initiating strand. The present study provides further evidence that in both three-strand and four-strand systems the pairing and strand exchange reactions share a common mechanism.     

Electron microscopic visualization of RecA-DNA filaments: Evidence for a cyclic extension of duplex DNA

As visualized by electron microscopy, RecA protein binds in a highly cooperative manner to single-stranded fd DNA in solutions of 0.01 M Tris (pH 7.5). The resulting nucleoprotein filament loops are 1.25 μm in length, have a fiber diameter of 12 nm and show an indication of a 4.5 nm repeat along the axis of the compact fibers. RecA binds to linear duplex fd DNA in solutions of 0.01 M Tris (pH 7.5) to yield chains of beads which, in the presence of Mg²⁺ and ATP, coalesce into smooth filaments with a length of 1.9 μm (the length of protein-free fd duplex DNA) and have a fiber diameter of 12 nm. In solutions containing Mg²⁺ and ATP-γ-S, however, RecA binds to duplex DNA in a highly cooperative manner to yield rigid filaments 3.0 μm in length. These filaments are 12 nm in diameter and show a very clear 7.5 nm axial repeat. This extension of DNA to 150% of its usual length in the apparent absence of any single-stranded components suggests that the DNA helix must also be highly unwound and provides new insights into the mode of RecA action.     

Assembly and disassembly of RecA protein filaments occur at opposite filament ends. Relationship to DNA strand exchange

RecA protein primarily associates with and dissociates from opposite ends of nucleoprotein filaments formed on linear duplex DNA. RecA nucleoprotein filaments that are hydrolyzing ATP therefore engage in a dynamic process under some conditions that has some of the properties of treadmilling. We have also investigated whether the net polymerization of recA protein at one end of the filament and/or a net depolymerization at the other end drives unidirectional strand exchange. There is no demonstrable correlation between recA protein association/dissociation and the strand exchange reaction. RecA protein-mediated DNA strand exchange is affected minimally by changes in reaction conditions (dilution, pH shift, or addition of small amounts of adenosine-5'-O-(3-thiotriphosphate) that have large and demonstrable effects on recA protein association, dissociation, or both. Rather than driving strand exchange, these assembly and disassembly processes may simply represent the mechanism by which recA nucleoprotein filaments are recycled in the cell.     

Investigating Deinococcus radiodurans RecA protein filament formation on double-stranded DNA by a real-time single-molecule approach

With the aid of an efficient, precise, and almost error-free DNA repair system, Deinococcus radiodurans can survive hundreds of double-strand breaks inflicted by high doses of irradiation or desiccation. RecA of D. radiodurans (DrRecA) plays a central role both in the early phase of repair by an extended synthesis-dependent strand annealing process and in the later more general homologous recombination phase. Both roles likely require DrRecA filament formation on duplex DNA. We have developed single-molecule tethered particle motion experiments to study the assembly dynamics of RecA proteins on individual duplex DNA molecules by observing changes in DNA tether length resulting from RecA binding. We demonstrate that DrRecA nucleation on double-stranded DNA is much faster than that of Escherichia coli RecA protein (EcRecA), but the extension is slower. This combination of attributes would tend to increase the number and decrease the length of DrRecA filaments relative to those of EcRecA, a feature that may reflect the requirement to repair hundreds of genomic double-strand breaks concurrently in irradiated Deinococcus cells.     

Synapsis and the formation of paranemic joints by E. coli RecA protein

E. coli RecA protein promotes the homologous pairing of a single strand with duplex DNA even when certain features of the substrates, such as circularity, prohibit the true intertwining of the newly paired strands. The formation of such nonintertwined or paranemic joints does not require superhelicity, and indeed can occur with relaxed closed circular DNA. E. coli topoisomerase I can intertwine the incoming single strand in the paranemic joint with its complement, thereby topologically linking single-stranded DNA to all of the duplex molecules in the reaction mixture. The efficiency of formation of paranemic joints, the time course, and estimates of their length, all suggest that they represent true synaptic intermediates in the pairing reaction promoted by RecA protein.     

RecA protein and SOS. Correlation of mutagenesis phenotype with binding of mutant RecA proteins to duplex DNA and LexA cleavage

The RecA protein of Escherichia coli is required for SOS-induced mutagenesis in addition to its recombinational and regulatory roles. We have suggested that RecA might participate directly in targeted mutagenesis by binding preferentially to the site of the DNA damage (e.g. pyrimidine dimer) because of its partially unwound nature; DNA polymerase III will then encounter RecA-coated DNA at the lesion and might replicate across the damaged site more often but with reduced fidelity. In support of this proposal, we have found that the phenotype of wild-type and mutant RecA for mutagenesis correlates with capacity to bind to double-stranded DNA. Wild-type RecA binds more efficiently to ultraviolet (u.v.)-irradiated, duplex DNA than to non-irradiated DNA. The RecA441 (Tif) protein that is constitutive for mutagenesis binds extremely well to double-stranded DNA with no lesions, whereas the RecA430 protein that is defective in mutagenesis binds poorly even to u.v.-irradiated DNA. The RecA phenotype also correlates with capacity to use duplex DNA as a cofactor for cleavage of the LexA repressor protein for SOS-controlled operons. Wild-type RecA provides efficient cleavage of LexA only with u.v.-irradiated duplex DNA; RecA441 cleaves well with non-irradiated DNA; RecA430 gives very poor cleavage even with u.v.-irradiated DNA. We conclude that the interaction of RecA with damaged double-stranded DNA is likely to be a critical component of SOS mutagenesis and to define a pathway for the LexA cleavage reaction as well.     

Fluorescence energy transfer in one dimension: Frequency-domain fluorescence study of DNA–fluorophore complexes

The theory for the salt dependence of the free energy, entropy, and enthalpy of a polyelectrolyte in the PB (PB) model is extended to treat the nonspecific salt dependence of polyelectrolyte–ligand binding reactions. The salt dependence of the binding constant (K) is given by the difference in osmotic pressure terms between the react ants and the products. For simple 1-1 salts it is shown that this treatment is equivalent to the general preferential interaction model for the salt dependence of binding [C. Anderson and M. Record (1993) Journal of Physical Chemistry, Vol. 97, pp. 7116–7126]. The salt dependence, entropy, and enthalpy are compared for the PB model and one specific form of the preferential interaction coefficient model that uses counterion condensation/limiting law (LL) behavior. The PB and LL models are applied to three ligand–polyelectrolyte systems with the same net ligand charge: a model sphere–cylinder binding reaction, a drug–DNA binding reaction, and a protein–DNA binding reaction. For the small ligands both the PB and limiting law models give (ln K vs. In [salt]) slopes close in magnitude to the net ligand charge. However, the enthalpy/entropy breakdown of the salt dependence is quite different. In the PB model there are considerable contributions from electrostatic enthalpy and dielectric (water reorientation) entropy, compared to the predominant ion cratic (release) entropy in the limiting law model. The relative contributions of these three terms in the PB model depends on the ligand: for the protein, ion release entropy is the smallest contribution to the salt dependence of binding. The effect of three approximations made in the LL model is examined: These approximations are (1) the ligand behaves ideally, (2) the preferential interaction coefficient of the polyelectrolyte is unchanged upon ligand binding, and (3) the polyelectrolyte preferential interaction coefficient is given by the limiting law/counterion-condensation value. Analysis of the PB model shows that assumptions 2 and 3 break down at finite salt concentrations. For the small ligands the effects on the slope cancel, however, giving net slopes that are similar in the PB and LL models, but with a different entropy/enthalpy breakdown. For the protein ligand the errors from assumptions 2 and 3 in the LL model do not cancel. In addition, the ligand no longer behaves ideally due to its complex structure and charge distribution. Thus for the protein the slope is no longer related simply to the net ligand charge, and the PB model gives a much larger slope than the LL model. Additionally, in the PB model most of the salt dependence of the protein binding comes from the change in ligand activity, i.e. from nonspecific anion effects, in contrast to the small ligand case. While the absolute binding is sensitive to polyelectrolyte length, little length effect is seen on the salt dependence for the small ligands at 0.1M salt, and for lengths > 60 Å. Almost no DNA length dependenceis seen in the salt dependence of the protein binding, since this is determined primarily by the protein, not the DNA. © 1995 John Wiley & Sons, Inc.     

Role of SSB protein in RecA promoted branch migration reactions

Summary The RecA protein ofEscherichia coli is essential for genetic recombination and postreplicational repair of DNA. In vitro, RecA protein promotes strand transfer reactions between full length linear duplex and single stranded circular DNA of X174 to form heteroduplex replicative form II-like structures (Cox and Lehman 1981a). In a similar way, it transfers one strand of a short duplex restriction fragment to a single stranded circle. Both reactions require RecA and single strand binding protein (SSB) in amounts sufficient to saturate the ssDNA. The rate and extent of strand transfer is enhanced considerably when SSB is added after preincubation of the DNA with RecA protein. In contrast, SSB protein is not required for RecA protein catalysed reciprocal strand exchanges between regions of duplex DNA. These results indicate that while SSB is necessary for efficient transfer between linear duplex and ssDNA to form a single heteroduplex, it is not required for branch migration reactions between duplex molecules that form two heteroduplexes.Abbreviations SSB single strand binding protein - ssDNA single stranded DNA - X phage X174 - bp base pairs - ATP[S] adenosine 5-O-(gamma-thiotriphosphate)     

Branch migration of three-strand recombination intermediates by RecG, a possible pathway for securing exchanges initiated by 3''-tailed duplex DNA.

RecG protein is required for normal levels of recombination and DNA repair in Escherichia coli. This 76 kDa polypeptide is a junction-specific DNA helicase that acts post-synaptically to drive branch migration of Holliday junction intermediates made by RecA during the strand exchange stage of recombination. To gain further insight into the role of RecG, we studied its activity on three-strand intermediates formed by RecA between circular single-stranded and linear duplex DNAs. Once RecA is removed, RecG drives branch migration of these intermediates by a junction-targeted activity that depends on hydrolysis of ATP. RuvAB has a similar activity. However, when RecG is added to a RecA strand exchange reaction it severely reduces the accumulation of joint molecule intermediates by driving branch migration of junctions in the reverse direction to that catalysed by RecA strand exchange. In comparison, RuvAB has little effect on the reaction. We discuss how reverse branch migration by RecG, which acts counter of the 5'-->3' polarity of RecA binding and strand exchange, could serve to promote or abort the early stages of recombination, depending on the orientation of the single DNA strand initiating the exchange relative to the adjacent duplex region.     

A DNA pairing-enhanced conformation of bacterial RecA proteins

The RecA proteins of Escherichia coli (Ec) and Deinococcus radiodurans (Dr) both promote a DNA strand exchange reaction involving two duplex DNAs. The four-strand exchange reaction promoted by the DrRecA protein is similar to that promoted by EcRecA, except that key parts of the reaction are inhibited by Ec single-stranded DNA-binding protein (SSB). In the absence of SSB, the initiation of strand exchange is greatly enhanced by dsDNA-ssDNA junctions at the ends of DNA gaps. This same trend is seen with the EcRecA protein. The results lead to an expansion of published hypotheses for the pathway for RecA-mediated DNA pairing, in which the slow first order step (observed in several studies) involves a structural transition to a state we designate P. The P state is identical to the state found when RecA is bound to double-stranded (ds) DNA. The structural state present when the RecA protein is bound to single-stranded (ss) DNA is designated A. The DNA pairing model in turn facilitates an articulation of three additional conclusions arising from the present work. 1) When a segment of a RecA filament bound to ssDNA is forced into the P state (as RecA bound to the ssDNA immediately adjacent to dsDNA-ssDNA junction), the segment becomes "pairing enhanced." 2) The unusual DNA pairing properties of the D. radiodurans RecA protein can be explained by postulating this protein has a more stringent requirement to initiate DNA strand exchange from the P state. 3) RecA filaments bound to dsDNA (P state) have directly observable structural changes relative to RecA filaments bound to ssDNA (A state), involving the C-terminal domain.     

RecA Protein from the extremely radioresistant bacterium Deinococcus radiodurans: expression, purification, and characterization

The RecA protein of Deinococcus radiodurans (RecA(Dr)) is essential for the extreme radiation resistance of this organism. The RecA(Dr) protein has been cloned and expressed in Escherichia coli and purified from this host. In some respects, the RecA(Dr) protein and the E. coli RecA (RecA(Ec)) proteins are close functional homologues. RecA(Dr) forms filaments on single-stranded DNA (ssDNA) that are similar to those formed by the RecA(Ec). The RecA(Dr) protein hydrolyzes ATP and dATP and promotes DNA strand exchange reactions. DNA strand exchange is greatly facilitated by the E. coli SSB protein. As is the case with the E. coli RecA protein, the use of dATP as a cofactor permits more facile displacement of bound SSB protein from ssDNA. However, there are important differences as well. The RecA(Dr) protein promotes ATP- and dATP-dependent reactions with distinctly different pH profiles. Although dATP is hydrolyzed at approximately the same rate at pHs 7.5 and 8.1, dATP supports an efficient DNA strand exchange only at pH 8.1. At both pHs, ATP supports efficient DNA strand exchange through heterologous insertions but dATP does not. Thus, dATP enhances the binding of RecA(Dr) protein to ssDNA and the displacement of ssDNA binding protein, but the hydrolysis of dATP is poorly coupled to DNA strand exchange. The RecA(Dr) protein thus may offer new insights into the role of ATP hydrolysis in the DNA strand exchange reactions promoted by the bacterial RecA proteins. In addition, the RecA(Dr) protein binds much better to duplex DNA than the RecA(Ec) protein, binding preferentially to double-stranded DNA (dsDNA) even when ssDNA is present in the solutions. This may be of significance in the pathways for dsDNA break repair in Deinococcus.     

RecA protein reinitiates strand exchange on isolated protein-free DNA intermediates. An ADP-resistant process

Efficient homologous pairing de novo of linear duplex DNA with a circular single strand (plus strand) coated with RecA protein requires saturation and extension of the single strand by the protein. However, strand exchange, the transfer of a strand from duplex DNA to the nucleoprotein filament, which follows homologous pairing, does not require the stable binding of RecA protein to single-stranded DNA. When RecA protein was added back to isolated protein-free DNA intermediates in the presence of sufficient ADP to inhibit strongly the binding of RecA protein to single-stranded DNA, strand exchange nonetheless resumed at the original rate and went to completion. Characterization of the protein-free DNA intermediate suggested that it has a special site or region to which RecA protein binds. Part of the nascent displaced plus strand of the deproteinized intermediate was unavailable as a cofactor for the ATPase activity of RecA protein, and about 30% resisted digestion by P1 endonuclease, which acts preferentially on single-stranded DNA. At the completion of strand exchange, when the distal 5' end of the linear minus strand had been fully incorporated into heteroduplex DNA, a nucleoprotein complex remained that contained all three strands of DNA from which the nascent displaced strand dissociated only over the next 50 to 60 minutes. Deproteinization of this intermediate yielded a complex that also contained three strands of DNA in which the nascent displaced strand was partially resistant to both Escherichia coli exonuclease I and P1 endonuclease. The deproteinized complex showed a broad melting transition between 37 degrees C and temperatures high enough to melt duplex DNA. These results show that strand exchange can be subdivided into two stages: (1) the exchange of base-pairs, which creates a new heteroduplex pair in place of a parental pair; and (2) strand separation, which is the physical displacement of the unpaired strand from the nucleoprotein filament. Between the creation of new heteroduplex DNA and the eventual separation of a third strand, there exists an unusual DNA intermediate that may contain three-stranded regions of natural DNA that are several thousand bases in length.