RecA-SSB Interaction Modulates RecA Nucleoprotein Filament Formation on SSB-Wrapped DNA

Abstract

E. coli RecA recombinase catalyzes the homology pairing and strand exchange reactions in homologous recombinational repair. RecA must compete with single-stranded DNA binding proteins (SSB) for single-stranded DNA (ssDNA) substrates to form RecA nucleoprotein filaments, as the first step of this repair process. It has been suggested that RecA filaments assemble mainly by binding and extending onto the free ssDNA region not covered by SSB, or are assisted by mediators. Using the tethered particle motion (TPM) technique, we monitored individual RecA filament assembly on SSB-wrapped ssDNA in real-time. Nucleation times of the RecA E38K nucleoprotein filament assembly showed no apparent dependence among DNA substrates with various ssDNA gap lengths (from 60 to 100 nucleotides) wrapped by one SSB in the (SSB)₆₅ binding mode. Our data have shown an unexpected RecA filament assembly mechanism in which a RecA-SSB-ssDNA interaction exists. Four additional pieces of evidence support our claim: the nucleation times of the RecA assembly varied (1) when DNA substrates contained different numbers of bound SSB tetramers; (2) when the SSB wrapping mode conversion is induced; (3) when SSB C-terminus truncation mutants are used; and (4) when an excess of C-terminal peptide of SSB is present. Thus, a RecA-SSB interaction should be included in discussing RecA regulatory mechanism.

Figure 1

SSB inhibits RecA nucleoprotein filament assembly. (A–C) Schematic illustration of the experimental setup and representative bead BM time-courses of RecA E38K assembly on various DNA substrates: (A) (dT)₃₅; (B) (dT)₁₀₀ and (C) (dT)₁₀₀ wrapped with a SSB tetramer. The reaction condition prepared the (SSB)₆₅ binding mode as previously documented. RecA filament assembly leads to the apparent bead BM increase. (D) Mean nucleation time (second) and (E) mean extension time (second/RecA) of RecA E38K assembling on these DNA substrates. Error bar is one standard error of the mean.

Figure 2

No apparent ssDNA length dependence of RecA E38K filament assembly kinetics on single SSB-wrapped ssDNA. (A) Mean nucleation time of RecA E38K on SSB-wrapped ssDNA-gap substrates ((dT)_n, n = 60–100) shows no apparent difference on ssDNA lengths (filled blue circles). RecA E38K nucleation times on SSB-free substrates are significantly shorter (<100 seconds, gray open squares). The red dashed line is the prediction based on a pure passive nucleation mechanism with τ ⁻¹ = k(L − 65) + C, where τ, k, L, 65 and C are mean nucleation time, microscopic rate constant, ssDNA gap length and number of nucleotides occupied by SSB in (SSB)₆₅ binding mode, constant, respectively. Reactions were done in the presence of 2 μM E38K RecA. Using lower concentrations of E38K RecA (0.18 μM, blue open circles) leads to longer nucleation time, but also shows no ssDNA-length-dependence. (B) RecA mean extension time (sec/RecA) did not show variation with ssDNA length. In the presence of SSB, the extension time is slower. (C) Total bead BM change upon RecA filament assembly followed the expected trend. Lines are drawn for guidance purpose. Error bar is one standard error of mean.

Figure 3

RecA E38K filament assembly dynamics vary with numbers of SSB wrapped on ssDNA. Mean nucleation time (A), histograms of nucleation time (B), mean extension time (C) and mean bead BM change (D) of RecA filament assembly on (dT)₇₀, (dT)₁₃₅, (dT)₂₀₀ and AC264 gapped DNA substrates. The reaction condition used results in SSB in (SSB)₆₅ binding mode, leading to 1–4 SSB bound in these substrates. Error bar is one standard error of the mean. Red lines in (B) are exponential fits.

Figure 4

Modifications to SSB alter RecA E38K filament assembly dynamics. (A) Percentage of the extended RecA E38K filament in the presence of bound wtSSB (WT) or bound SSBΔC8 after 5 minutes and 20 minutes of RecA E38K addition (see Figure S5). (B–D) Mean nucleation time (B), mean extension time (C) and bead BM change (D) of RecA E38K on ssDNA substrates associated with various SSB proteins and mutants. (E–G) Mean nucleation time (E), mean extension time (F) and bead BM change (G) of RecA E38K incubated with 2 μM SSB C-terminal peptides (wild-type or scrambled). The E38K-C peptide mixture was flowed into the reaction chamber containing the SSB-wrapped ssDNA substrates. Pre-incubation with the wild-type C-peptide of SSB significantly delays nucleation and extension of RecA E38K, compared to the experiments using scrambled C-peptide of SSB. All assembly kinetics were determined using (dT)₉₀ gapped ssDNA substrates, associated with either no SSB, bound wtSSB, bound wtSSB plus free wtSSB, bound SSBΔC8, bound SSB113 or with bound SSBΔF. Error bar is one standard error of the mean.

Figure 5

A model of RecA filament assembly on SSB-wrapped ssDNA substrates. In the beginning, SSB wraps around ssDNA tail with high affinity (Top). In case there is only singly bound SSB in its (SSB)₆₅ binding mode (left panel), RecA has to either interact with the C-terminus of SSB or directly assemble on exposed ssDNA region to form a stable nucleus, followed by fast extension and complete removal of SSB. Once ssDNA is bound by multiple numbers of SSB (middle panel), RecA spends longer time nucleating and extending on ssDNA, especially for even-numbered SSB DNA substrate. RecA is incapable of displacing entire SSB on ssDNA. In the presence of excess SSB (right panel), SSB prefers (SSB)₃₅ binding mode; the highly cooperative nature of (SSB)₃₅ binding mode forces RecA to nucleate in a slower mechanism, followed by slower extension and incomplete removal of SSB.