Stabilization of recA protein-ssDNA complexes by the single-stranded DNA binding protein of Escherichia coli

In vitro recombination reactions promoted by the recA protein of Escherichia coli are enhanced by the single-stranded DNA binding protein (SSB). SSB affects the assembly of the filamentous complexes between recA protein and ssDNA that are the active form of the recA protein. Here, we present evidence that SSB plays a complex role in maintaining the stability and activity of recA-ssDNA filaments. Results of ATPase, nuclease protection, and DNA strand exchange assays suggest that the continuous presence of SSB is required to maintain the stability of recA-ssDNA complexes under reaction conditions that support their recombination activity. We also report data that indicate that there is a functional distinction between the species of SSB present at 10 mM magnesium chloride, which enhances recA-ssDNA binding, and a species present at 1 mM magnesium chloride, which displaces recA protein from ssDNA. These results are discussed in the context of current models of SSB conformation and of SSB action in recombination activities of the recA protein.     

Assembly of presynaptic filaments. Factors affecting the assembly of RecA protein onto single-stranded DNA

We have previously shown that the assembly of RecA protein onto single-stranded DNA (ssDNA) facilitated by SSB protein occurs in three steps: (1) rapid binding of SSB protein to the ssDNA; (2) nucleation of RecA protein onto this template; and (3) co-operative polymerization of additional RecA protein to yield presynaptic filaments. Here, electron microscopy has been used to further explore the parameters of this assembly process. The optimal extent of presynaptic filament formation required at least one RecA protein monomer per three nucleotides, high concentrations of ATP (greater than 3 mM in the presence of 12 mM-Mg2+), and relatively low concentrations of SSB protein (1 monomer per 18 nucleotides). Assembly was depressed threefold when SSB protein was added to one monomer per nine nucleotides. These effects appeared to be exerted at the nucleation step. Following nucleation, RecA protein assembled onto ssDNA at net rates that varied from 250 to 900 RecA protein monomers per minute, with the rate inversely related to the concentration of SSB protein. Combined sucrose sedimentation and electron microscope analysis established that SSB protein was displaced from the ssDNA during RecA protein assembly.     

recA protein promoted DNA strand exchange

recA protein and circular single-stranded DNA form a stable complex in the presence of single-stranded DNA binding protein (SSB), in which one recA protein monomer is bound per two nucleotides of DNA. These complexes are kinetically significant intermediates in the exchange of strands between the single-stranded DNA and an homologous linear duplex. After completion of strand exchange, the recA protein remains tightly associated with the circular duplex product of the reaction and the SSB is bound to the displaced linear single strand. Upon addition of ADP, the recA protein-duplex DNA complex dissociates. RecA protein also interacts with single-stranded DNA in the absence of SSB; however, the amount of recA protein bound is substantially reduced. These findings provide direct physical evidence for the participation of SSB in the formation of the recA protein-single-stranded DNA complexes inferred earlier from kinetic analysis. Moreover, they confirm the ability of recA protein to equilibrate between bound and free forms in the absence of SSB.     

High-resolution AFM imaging of single-stranded DNA-binding (SSB) protein--DNA complexes

DNA in living cells is generally processed via the generation and the protection of single-stranded DNA involving the binding of ssDNA-binding proteins (SSBs). The studies of SSB-binding mode transition and cooperativity are therefore critical to many cellular processes like DNA repair and replication. However, only a few atomic force microscopy (AFM) investigations of ssDNA nucleoprotein filaments have been conducted so far. The point is that adsorption of ssDN A-SSB complexes on mica, necessary for AFM imaging, is not an easy task. Here, we addressed this issue by using spermidine as a binding agent. This trivalent cation induces a stronger adsorption on mica than divalent cations, which are commonly used by AFM users but are ineffective in the adsorption of ssDNA-SSB complexes. At low spermidine concentration (<0.3 mM), we obtained AFM images of ssDNA-SSB complexes (E. coli SSB, gp32 and yRPA) on mica at both low and high ionic strengths. In addition, partially or fully saturated nucleoprotein filaments were studied at various monovalent salt concentrations thus allowing the observation of SSB-binding mode transition. In association with conventional biochemical techniques, this work should make it possible to study the dynamics of DNA processes involving DNA-SSB complexes as intermediates by AFM.     

Continuous association of Escherichia coli single-stranded DNA binding protein with stable complexes of recA protein and single-stranded DNA

The single-stranded DNA binding protein of Escherichia coli (SSB) stimulates recA protein promoted DNA strand exchange reactions by promoting and stabilizing the interaction between recA protein and single-stranded DNA (ssDNA). Utilizing the intrinsic tryptophan fluorescence of SSB, an ATP-dependent interaction has been detected between SSB and recA-ssDNA complexes. This interaction is continuous for periods exceeding 1 h under conditions that are optimal for DNA strand exchange. Our data suggest that this interaction does not involve significant displacement of recA protein in the complex by SSB when ATP is present. The properties of this interaction are consistent with the properties of SSB-stabilized recA-ssDNA complexes determined by other methods. The data are incompatible with models in which SSB is displaced after functioning transiently in the formation of recA-ssDNA complexes. A continuous association of SSB with recA-ssDNA complexes may therefore be an important feature of the mechanism by which SSB stimulates recA protein promoted reactions.     

Escherichia coli single-strand binding protein forms multiple, distinct complexes with single-stranded DNA

Four distinct binding modes for the interaction of Escherichia coli single-strand binding (SSB) protein with single-stranded (ss) DNA have been identified on the basis of quantitative titrations that monitor the quenching of the SSB protein fluorescence upon binding to the homopolynucleotide poly(dT) over a range of MgCl2 and NaCl concentrations at 25 and 37 degrees C. This is the first observation of multiple binding modes for a single protein binding to DNA. These results extend previous studies performed in NaCl (25 degrees C, pH 8.1), in which two distinct SSB-ss DNA binding modes possessing site sizes of 33 and 65 nucleotides per bound SSB tetramer were observed [Lohman, T.M., & Overman, L. B. (1985) J. Biol. Chem. 260, 3594-3603]. Each of these binding modes differs in the number of nucleotides occluded upon interaction with ss DNA (i.e., site size). Along with the previously observed modes with site sizes of 35 +/- 2 and 65 +/- 3 nucleotides per tetramer, a third distinct binding mode, at 25 degrees C, has been identified, possessing a site size of 56 +/- 3 nucleotides per bound SSB tetramer, which is stable over a wide range of MgCl2 concentrations. At 37 degrees C, a fourth binding mode is observed, possessing a site size of 40 +/- 2 nucleotides per tetramer, although this mode is observable only over a small range of salt concentration.(ABSTRACT TRUNCATED AT 250 WORDS)     

Interaction of the recA protein of Escherichia coli with single-stranded DNA

The interaction of recA protein with single-stranded (ss) phi X174 DNA has been examined by means of a nuclease protection assay. The stoichiometry of protection was found to be 1 recA monomer/approximately 4 nucleotides of ssDNA both in the absence of a nucleotide cofactor and in the presence of ATP. In contrast, in the presence of adenosine 5'-O-(thiotriphosphate) (ATP gamma S) the stoichiometry was 1 recA monomer/approximately 8 nucleotides. No protection was seen with ADP. In the absence of a nucleotide cofactor, the binding of recA protein to ssDNA was quite stable as judged by equilibration with a challenge DNA (t1/2 approximately 30 min). Addition of ATP stimulated this transfer (t1/2 approximately 3 min) as did ADP (t1/2 approximately 0.2 min). ATP gamma S greatly reduced the rate of equilibration (t1/2 greater than 12 h). Direct visualization of recA X ssDNA complexes at subsaturating recA protein concentrations using electron microscopy revealed individual ssDNA molecules partially covered with recA protein which were converted to highly condensed networks upon addition of ATP gamma S. These results have led to a general model for the interaction of recA protein with ssDNA.     

Biochemical properties of the Escherichia coli recA430 protein. Analysis of a mutation that affects the interaction of the ATP-recA protein complex with single-stranded DNA

The biochemical properties of the recA430 protein have been examined and compared to those of wild-type recA protein. We find that, while the recA430 protein possesses ssDNA-dependent rATP activity, this activity is inhibited by the Escherichia coli single-stranded DNA binding protein (SSB protein) under many conditions that enhance wild-type recA protein rATPase hydrolysis. Stimulation of rATPase activity by SSB protein is observed only at high concentrations of both rATP (greater than 1 mM) and recA430 protein (greater than 5 microM). In contrast, stimulation of ssDNA-dependent dATPase activity by SSB protein is less sensitive to protein and nucleotide concentration. Consistent with the nucleotide hydrolysis data, recA430 protein can carry out DNA strand exchange in the presence of either rATP or dATP. However, in the presence of rATP, both the rate and the extent of DNA strand exchange by recA430 protein are greatly reduced compared to wild-type recA protein and are sensitive to recA430 protein concentration. This reduction is presumably due to the inability of recA430 protein to compete with SSB protein for ssDNA binding sites under these conditions. The cleavage of lexA repressor protein by recA430 protein is also sensitive to the nucleotide cofactor present and is completely inhibited by SSB protein when rATP is the cofactor but not when dATP is used. Finally, the steady-state affinity and the rate of association of the recA430 protein-ssDNA complex are reduced, suggesting that the mutation affects the interaction of the ATP-bound form of recA protein with ssDNA. This alteration is the likely molecular defect responsible for inhibition of recA430 protein rATP-dependent function by SSB protein. The biochemical properties observed in the presence of dATP and SSB protein, i.e. the reduced levels of both DNA strand exchange activity and cleavage of lexA repressor protein, are consistent with the phenotypic behavior of recA430 mutations.     

Effects of Escherichia coli SSB protein on the single-stranded DNA-dependent ATPase activity of Escherichia coli RecA protein: Evidence that SSB protein facilitates the binding of RecA protein to regions of secondary structure within single-stranded DNA

The effect that Escherichia coli single-stranded DNA binding (SSB) protein has on the single-stranded DNA-dependent ATPase activity of RecA protein is shown to depend upon a number of variables such as order of addition, magnesium concentration, temperature and the type of single-stranded DNA substrate used. When SSB protein is added to the DNA solution prior to the addition of RecA protein, a significant inhibition of ATPase activity is observed. Also, when SSB protein is added after the formation of a RecA protein-single-stranded DNA complex using either etheno M13 DNA, poly(dA) or poly(dT), or using single-stranded phage M13 DNA at lower temperature (25 °C) and magnesium chloride concentrations of 1 mm or 4 mm, a time-dependent inhibition of activity is observed. These results are consistent with the conclusion that SSB protein displaces the RecA protein from these DNA substrates, as described in the accompanying paper. However, if SSB protein is added last to complexes of RecA protein and single-stranded M13 DNA at elevated temperature (37 °C) and magnesium chloride concentrations of 4 mm or 10 mm, or to poly(dA) and poly(dT) that was renatured in the presence of RecA protein, no inhibition of ATPase activity is observed; in fact, a marked stimulation is observed for single-stranded M13 DNA. A similar effect is observed if the bacteriophage T4-coded gene 32 protein is substituted for SSB protein. The apparent stoichiometry of DNA (nucleotides) to RecA protein at the optimal ATPase activity for etheno M13 DNA, poly(dA) and poly(dT) is 6(±1) nucleotides per RecA protein monomer at 4 mm-MgCl₂ and 37 °C. Under the same conditions, the apparent stoichiometry obtained using single-stranded M13 DNA is 12 nucleotides per RecA protein monomer; however, the stoichiometry changes to 4.5 nucleotides per RecA protein monomer when SSB protein is added last. In addition, a stoichiometry of four nucleotides per RecA protein can be obtained with single-stranded M13 DNA in the absence of SSB protein if the reactions are carried out in 1 mm-MgCl₂. These data are consistent with the interpretation that secondary structure within the natural DNA substrate limits the accessibility of RecA protein to these regions. The role of SSB protein is to eliminate this secondary structure and allow RecA protein to bind to these previously inaccessible regions of the DNA. In addition, our results have disclosed an additional property of the RecA protein-single-stranded DNA complex: namely, in the presence of complementary base-pairing and at elevated temperatures and magnesium concentrations, a unique RecA protein-DNA complex forms that is resistant to inhibition by SSB protein.     

Formation of D loops by the UvsX protein of T4 bacteriophage: a comparison of the reaction catalyzed in the presence or absence of gene 32 protein

The UvsX protein of T4 bacteriophage will catalyze the formation of D loops between linear single-stranded DNA (ssDNA) and homologous supercoiled double-stranded DNA (dsDNA) in the absence of T4 gene 32 protein (gp32). This reaction requires one monomer of UvsX protein per three nucleotides of ssDNA so that the ssDNA is completely covered with UvsX protein. Under these conditions, high rates of ATP hydrolysis are observed, and one-third of the products are joined paranemically. The reaction proceeds through a mechanism that creates homology-independent coaggregates of UvsX protein, dsDNA, and ssDNA. When UvsX protein is added to only 1 monomer per 8 nucleotides, but with 1 monomer of gp32 per 12 nucleotides, the rate of ATP hydrolysis is depressed, but D-loop formation is enhanced. Nearly all of the product is bound in plectonemic joints, and no coaggregated intermediates are formed. Coaggregate formation at high concentrations of UvsX protein is not inhibited by the presence of gp32; gp32 simply allows for efficient formation of D loops at such low concentrations of UvsX protein that coaggregates are not constructed. Electron microscopic visualization of the joint structures in this reaction reveals that both gp32 and UvsX protein are bound to the ssDNA. The single-stranded DNA binding (SSB) protein of Escherichia coli will substitute only partially for gp32: in the presence of SSB protein, D-loop formation can be catalyzed at one UvsX protein monomer per eight nucleotides, and it is accomplished without the formation of coaggregates, but a major portion of the product is joined paranemically.     

Force and ATP hydrolysis dependent regulation of RecA nucleoprotein filament by single-stranded DNA binding protein

In Escherichia coli, the filament of RecA formed on single-stranded DNA (ssDNA) is essential for recombinational DNA repair. Although ssDNA-binding protein (SSB) plays a complicated role in RecA reactions in vivo, much of our understanding of the mechanism is based on RecA binding directly to ssDNA. Here we investigate the role of SSB in the regulation of RecA polymerization on ssDNA, based on the differential force responses of a single 576-nucleotide-long ssDNA associated with RecA and SSB. We find that SSB outcompetes higher concentrations of RecA, resulting in inhibition of RecA nucleation. In addition, we find that pre-formed RecA filaments de-polymerize at low force in an ATP hydrolysis- and SSB-dependent manner. At higher forces, re-polymerization takes place, which displaces SSB from ssDNA. These findings provide a physical picture of the competition between RecA and SSB under tension on the scale of the entire nucleoprotein SSB array, which have broad biological implications particularly with regard to competitive molecular binding.     

Investigation of binding between recA protein and single-stranded polynucleotides with the aid of a fluorescent deoxyribonucleic acid derivative

The availability of epsilon DNA, a fluorescent ssDNA derivative, has made it possible to examine quantitatively the interactions between recA protein and single-stranded polynucleotides. Fluorescence titrations of epsilon DNA with recA protein and vice versa establish that each recA protein monomer covers 5.5 epsilon DNA nucleotides and that the dissociation constant of the recA-epsilon DNA complex is 10 nM. Fluorescence titrations of recA protein-epsilon DNA mixtures with poly(dT) establish that each recA protein monomer covers 5.1 poly(dT) nucleotides and that the dissociation constant of the recA-poly(dT) complex is 0.03 nM. Observations on how the addition of ssDNA affects the fluorescence of recA protein-epsilon DNA mixtures establish that the dissociation constant of the recA-ssDNA complex exceeds 20 microM. Stopped-flow kinetics in which excess recA protein binds to epsilon DNA indicate that k2 = 6 X 10(6) M-1 s-1 for the process. A more approximate kinetic technique indicates that recA protein binds to epsilon DNA at least one-tenth as fast as to poly(dT); the rate constant for dissociation of recA-epsilon DNA exceeds that for recA-poly(dT) by at least 30-fold. epsilon DNA is proven to be a versatile reagent for studying single-stranded polynucleotide-protein interactions. Not only can its own complexes with protein be investigated but also, under suitable circumstances, it can be used as a fluorescent probe to explore complexes incorporating nonfluorescent polynucleotides.     

A postsynaptic role for single-stranded DNA-binding protein in recA protein-promoted DNA strand exchange.

The single-stranded DNA-binding protein (SSB protein) is required for efficient genetic recombination in vivo. One function for SSB protein in DNA strand exchange in vitro is to remove secondary structure from single-stranded DNA (ssDNA) and thereby aid in the formation of recA protein-saturated presynaptic complexes. In the preceding paper (Lavery, P. E., and Kowalczykowski, S. C. (1992) J. Biol. Chem. 267, 9307-9314) we demonstrated that DNA strand exchange can occur in the presence of volume-occupying agents at low magnesium ion concentration, where secondary structures are reduced. Our results suggest that SSB protein is not acting during presynapsis under these conditions, yet the DNA strand exchange reaction is stimulated by the addition of SSB protein. In this study we present biochemical evidence which suggests that SSB protein stimulates DNA strand exchange by binding to the ssDNA displaced from joint molecules, thereby stabilizing them and allowing branch migration to extend the region of heteroduplex DNA. Therefore, our results indicate dual roles for SSB protein at elevated magnesium ion concentration; it functions during presynapsis, removing secondary structure from ssDNA, as indicated previously, and it also functions postsynaptically, binding to the ssDNA displaced from joint molecules.     

SSB protein limits RecOR binding onto single-stranded DNA

The RecO and RecR proteins form a complex that promotes the nucleation of RecA protein filaments onto SSB protein-coated single-stranded DNA (ssDNA). However, even when RecO and RecR proteins are provided at optimal concentrations, the loading of RecA protein is surprisingly slow, typically proceeding with a lag of 10 min or more. The rate-limiting step in RecOR-promoted RecA nucleation is the binding of RecOR protein to ssDNA, which is inhibited by SSB protein despite the documented interaction between RecO and SSB. Full activity of RecOR is seen only when RecOR is preincubated with ssDNA prior to the addition of SSB. The slow binding of RecOR to SSB-coated ssDNA involves the C terminus of SSB. When an SSB variant that lacks the C-terminal 8 amino acids is used, the capacity of RecOR to facilitate RecA loading onto the ssDNA is largely abolished. The results are used in an expanded model for RecOR action.     

Biochemical basis of the constitutive repressor cleavage activity of recA730 protein. A comparison to recA441 and recA803 proteins.

The recA730 mutation results in constitutive SOS and prophage induction. We examined biochemical properties of recA730 protein in an effort to explain the constitutive activity observed in recA730 strains. We find that recA730 protein is more proficient than the wild-type recA protein in the competition with single-stranded DNA binding protein (SSB protein) for single-stranded DNA (ssDNA) binding sites. Because an increased aptitude in the competition with SSB protein has been previously reported for recA441 protein and recA803 protein, we directly compared their in vitro activities with those of recA730 protein. At low magnesium ion concentration, both ATP hydrolysis and lexA protein cleavage experiments demonstrate that these recA proteins displace SSB protein from ssDNA in a manner consistent with their in vivo repressor cleavage activity, i.e. recA730 protein > recA441 protein > recA803 protein > recAwt protein. Additionally, a correlation exists between the proficiency of the recA proteins in SSB protein displacement and their rate of association with ssDNA. We propose that an increased rate of association with ssDNA allows recA730 protein to displace SSB protein from the ssDNA that occurs naturally in Escherichia coli and thereby to become activated for the repressor cleavage that leads to SOS induction. RecA441 protein is similarly activated for repressor cleavage; however, in this case, significant SSB protein displacement occurs only at elevated temperature. At physiological magnesium ion concentration, we argue that recA803 protein and wild-type recA protein do not displace sufficient SSB protein from ssDNA to constitutively induce the SOS response.     

ATP-independent renaturation of complementary DNA strands by the mutant recA1 protein from Escherichia coli

In an effort to clarify the requirement for ATP in the recA protein-promoted renaturation of complementary DNA strands, we have analyzed the mutant recA1 protein which lacks single-stranded DNA-dependent ATPase activity at pH 7.5. Like the wild type, the recA1 protein binds to single-stranded DNA with a stoichiometry of one monomer per approximately four nucleotides. However, unlike the wild type, the mutant protein is dissociated from single-stranded DNA in the presence of ATP or ADP. The ATP analogue adenosine 5'-O-3' (thiotriphosphate) appears to stabilize the binding of recA1 protein to single-stranded DNA but does not elicit the stoichiometry of 1 monomer/8 nucleotides or the formation of highly condensed protein-DNA networks that are characteristic of the wild type recA protein in the presence of this analogue. The recA1 protein does not catalyze DNA renaturation in the presence of ATP, consistent with the dissociation of recA1 protein from single-stranded DNA under these conditions. However, it does promote a pattern of Mg2+-dependent renaturation identical to that found for wild type recA protein.     

The distribution of Escherichia coli recA protein bound to duplex DNA with single-stranded ends

A short single-stranded tail on one end of an otherwise duplex DNA molecule enables recA protein, in the presence of ATP and MgCl2, to form a complex with the DNA which extends into the duplex portion of the molecule. Nuclease protection studies at a concentration of MgCl2 which permits homologous pairing showed that cleavage by restriction endonucleases at sites throughout the duplex region was inhibited, whereas digestion by DNase I was not affected. These results indicate that recA protein binds to the duplex portion of tailed DNA allowing access by DNase I to a random sample of the many sites at which it cleaves, but providing limited protection of the relatively rare restriction sites. Electron microscopy revealed that the recA nucleoprotein complex with duplex DNA is indeed a segmented or interrupted filament that, with time, extends further from the single-stranded tail into the duplex region. recA protein binding extended into the duplex region more rapidly for duplexes with 5' tails than for those with 3' tails. These observations show that recA protein translocates from a single-stranded region into duplex DNA in the form of a segmented filament by a mechanism that is not strongly polarized.     

The RecOR proteins modulate RecA protein function at 5' ends of single-stranded DNA.

The Escherichia coli RecF, RecO and RecR pro teins have previously been implicated in bacterial recombinational DNA repair at DNA gaps. The RecOR-facilitated binding of RecA protein to single-stranded DNA (ssDNA) that is bound by single-stranded DNA-binding protein (SSB) is much faster if the ssDNA is linear, suggesting that a DNA end (rather than a gap) facilitates binding. In addition, the RecOR complex facilitates RecA protein-mediated D-loop formation at the 5' ends of linear ssDNAs. RecR protein remains associated with the RecA filament and its continued presence is required to prevent filament disassembly. RecF protein competes with RecO protein for RecR protein association and its addition destabilizes RecAOR filaments. An enhanced function of the RecO and RecR proteins can thus be seen in vitro at the 5' ends of linear ssDNA that is not as evident in DNA gaps. This function is countered by the RecF/RecO competition for association with the RecR protein.     

Molecular visualization of the yeast Dmc1 protein ring and Dmc1-ssDNA nucleoprotein complex

Saccharomyces cerevisiae Dmc1, a meiosis-specific homologue of RecA, catalyzes homologous pairing and strand exchange during meiotic DNA recombination. The purified budding yeast Dmc1 (ScDmc1) protein exhibits much weaker recombinase activity in vitro as compared to that of the Escherichia coli RecA protein. Using atomic force microscopy (AFM) with carbon nanotube tips, we found ScDmc1 forms rings with an external diameter of 18 nm and a central cavity of 4 nm. In the presence of single-stranded DNA (ssDNA), the majority of the ScDmc1 protein (90%) bound DNA as protein rings; only a small faction (10%) was able to form filamentous structure. In contrast, nearly all RecA proteins form fine helical nucleoprotein filaments with ssDNA under identical conditions. RecA-mediated recombinase activity is initiated through the nucleation of RecA onto ssDNA to form helical nucleoprotein filaments. Our results support the notion that ScDmc1 becomes catalytically active only when it forms a helical nucleoprotein filament with ssDNA.     

Homology-dependent changes in adenosine 5'-triphosphate hydrolysis during recA protein promoted DNA strand exchange: evidence for long paranemic complexes

As a first step in DNA strand exchange, recA protein forms a filamentous complex on single-stranded DNA (ssDNA), which contains stoichiometric (one recA monomer per four nucleotides) amounts of recA protein. recA protein monomers within this complex hydrolyze ATP with a turnover number of 25 min-1. Upon introduction of linear homologous duplex DNA to initiate strand exchange, this rate of ATP hydrolysis drops by 33%. The decrease in rate is complete in less than 2 min, and the rate of ATP hydrolysis then remains constant during and subsequent to the strand exchange reaction. This drop is completely dependent upon homology in the duplex DNA. In addition, the magnitude of the drop is linearly dependent upon the length of the homologous region in the linear duplex DNA. Linear DNA substrates in which pairing is topologically restricted to a paranemic joint also follow this relationship. Taken together, these properties imply that all of the available homology in the incoming duplex DNA is detected very early in the DNA strand exchange reaction, with the linear duplex DNA paired paranemically with the homologous ssDNA in the complex throughout its length. The results indicate that paranemic joints can extend over thousands of base pairs. We note elsewhere [Pugh, B. F., & Cox, M. M. (1987b) J. Biol. Chem. 262, 1337-1343] that this duplex acquires resistance to digestion by DNase with a much slower time course (30 min), which parallels the progress of strand exchange. Together these results imply that the duplex DNA is paired with the ssDNA but remains outside the nucleoprotein filament. Finally, the results also support the notion that ATP hydrolysis occurs throughout the recA nucleoprotein filament.