= Solution
\b[The central structural result is that strand exchange takes place on an ATP-stabilized helical protein–DNA scaffold that stretches DNA and presents its bases for complementary pairing.] <X-ray crystallography> established local geometry and nucleotide-dependent contacts; <electron microscopy>, especially <cryo-electron microscopy>, revealed extended assemblies and captured donor-<DNA>-bound and strand-exchanged states. Their combination explains much of the molecular recognition mechanism while leaving the temporal sequence to be tested kinetically.
In <homologous recombination>, a <RecA>-family <recombinase> assembles on <single-stranded DNA> to make a <presynaptic filament>. The assembly encounters <double-stranded DNA>, tests <nucleic acid complementarity> and promotes <DNA strand exchange>, generating a <heteroduplex DNA> region. Local invasion yields a <D-loop>, with one strand of the donor displaced. Bacterial <RecA> and eukaryotic <RAD51> are related but are not interchangeable descriptions of every regulatory step. In a cellular <DNA repair> reaction, production and protection of the initial <single-stranded DNA> and subsequent <DNA> synthesis require other factors; a structure of the recombinase core does not explain the whole pathway.
An early contribution of <electron microscopy> was to show that the active species is an extended <nucleoprotein filament>, rather than a single <protein> clamping two short sequences together. Helical reconstruction and comparisons across nucleotide conditions distinguish extended and more compact assemblies. This organization suggested collective control: neighboring subunits could coordinate <DNA> binding and <ATP> state. Low-resolution filament dimensions alone could not locate individual bases or prove which contacts recognize complementarity; those questions required higher-resolution structures and biochemical tests.
The <RecA>–<DNA> crystal structures made that geometry explicit. They showed <ATP>-state-dependent contacts across neighboring <RecA> subunits and <DNA> held inside the helical assembly. The <DNA> is globally extended and underwound, while short three-base groups retain approximately ordinary local base geometry. Separation is concentrated between these groups. Thus the <protein> reshapes the overall substrate without making every adjacent <base pair> equally distorted. The related <single-stranded DNA> and <heteroduplex DNA> structures further showed that the incoming complementary strand is supported substantially through base pairing, providing a structural explanation for sequence dependence.<Mechanism of homologous recombination from the RecA–ssDNA dsDNA structures by Zhucheng Chen and others>[Figs. 1–5, presynaptic and postsynaptic filaments, nucleotide interfaces and DNA triplet organization]{ref}
This geometry has several mechanistic consequences. Organizing bases into small groups makes a local complementarity test possible before a long donor region has been opened. The scaffold also exposes <DNA> bases while imposing an energetic and geometrical cost on bringing intact donor <double-stranded DNA> into the same arrangement. Favorable new <Watson-Crick base pairing> can compensate for disrupting old pairing. Because <RecA> mainly organizes the searching strand rather than reading a fixed base sequence with a sequence-specific pocket, the same <recombinase> can act at many genomic sequences. These are mechanistic interpretations of the structures, not direct measurements of search speed.
The nucleotide-binding site lies at the subunit interface, so <ATP> binding and <DNA> affinity are coupled to filament organization. A change in nucleotide state can consequently affect several neighboring contacts rather than only a local catalytic site. Distinguish two roles: <ATP> binding stabilizes a strand-exchange-competent arrangement, whereas <ATP hydrolysis> supports filament turnover and completion of physiological reactions. A structure trapped with a nucleotide analogue cannot establish that one <ATP> is hydrolyzed for every <base pair> exchanged. Testing that assertion would require simultaneous nucleotide-turnover and strand-exchange measurements, including reactions lacking complementarity.
<Crystallography> also imposes selection. A continuous filament is hard to crystallize in a way that resolves its <DNA>, so engineered finite assemblies and stabilized nucleotide conditions can be useful. Crystal contacts, linked subunits or trapped intermediates may restrict the accessible structures. The strength of the result lies in an explicit atomic arrangement that explains independent observations; the limitation is that a stable crystal state does not reveal the distribution or lifetime of all intermediates in solution. Cross-check filament geometry in <electron microscopy>, test activity of the crystallized construct, and mutate candidate contacts while measuring assembly as well as <DNA strand exchange>.
<cryo-electron microscopy> extended this interpretation to human <RAD51>. Structures of presynaptic and postsynaptic <RAD51>–<DNA> filaments demonstrated conservation of the central scaffold and also identified differences in subunit interfaces and <DNA> interactions. An arrested synaptic assembly supplied evidence about the stage in which the donor has encountered the searching filament.<Cryo-EM structures of human RAD51 recombinase filaments during catalysis of DNA-strand exchange by Jingfei Xu and others>[Figs. 1–5, comparison of presynaptic, postsynaptic and arrested synaptic complexes]{ref} This supports a common mechanistic framework across organisms without assuming identical nucleotide kinetics, regulatory partners or sequence discrimination. It is particularly useful to compare like nucleotide and substrate conditions, because structural differences can otherwise reflect preparation rather than a genuine species difference.
A more direct view of the donor encounter came from <cryo-electron microscopy> of <RecA> synaptic and <D-loop> assemblies. In those structures the <C-terminus>-containing <protein domain> of <RecA> recruits donor <double-stranded DNA>; a <DNA>-binding loop helps open the duplex. The displaced strand is accommodated at a secondary binding region while the complementary strand can pair with the initiating <single-stranded DNA>. Structures with limited complementarity locate the new <heteroduplex DNA> and its junctions to the remaining donor duplex.<Mechanism of strand exchange from RecA–DNA synaptic and D-loop structures by Haijuan Yang and others>[Figs. 1–4, nonhomologous synapses, donor opening and D-loop complexes]{ref} This supplies a physical route from donor capture to a local pairing test and then strand displacement, rather than inferring the entire reaction merely from its endpoints.
The general model is a series of reversible local tests. Nonmatching encounters should tend to close or detach before opening a long region; favorable complementarity should permit longer-lived opening and extension. This avoids requiring every genomic encounter to pay the energetic cost of unwinding a long donor segment. Multiple structures support that interpretation, but their existence alone does not establish which state precedes which, how rapidly they interconvert, or the probability of successful invasion in a <biological cell>. Population fractions after trapping can differ from flux through the working reaction.
A discriminating experimental programme would compare homologous and mismatched donors of the same length, an active versus altered donor-opening loop, and <ATP> versus an appropriate trapped nucleotide state. Measure <DNA> binding, filament formation, nucleotide turnover and <D-loop> yield separately. A donor-opening <mutation> that abolishes <D-loop> formation while retaining a normal <presynaptic filament> is stronger evidence for its proposed role than a <mutation> that simply unfolds the <protein>. Check structures of the altered complex where possible. Time-resolved <Förster resonance energy transfer> or other single-molecule assays can test donor opening and exchange, with labels positioned and controlled so that unwinding is not confused with aggregation or spontaneous strand displacement. \b[Structures identify the scaffold and candidate intermediates; complementary binding, kinetic and mutational experiments establish how that scaffold performs a productive homology search.]
Back to article page
