RecX protein abrogates ATP hydrolysis and strand exchange promoted by RecA: insights into negative regulation of homologous recombination

Abstract

In many eubacteria, coexpression of recX with recA is essential for attenuation of the deleterious effects of recA overexpression; however, the molecular mechanism has remained enigmatic. Here, we show that Mycobacterium tuberculosis RecX binds directly to M. tuberculosis RecA as well as M. smegmatis and E. coli RecA proteins in vivo and in vitro, but not single-stranded DNA binding protein. The direct association of RecX with RecA failed to regulate the specificity or extent of binding of RecA either to DNA or ATP, ligands that are central to activation of its functions. Significantly, RecX severely impeded ATP hydrolysis and the generation of heteroduplex DNA promoted by homologous, as well as heterologous, RecA proteins. These findings reveal a mode of negative regulation of RecA, and imply that RecX might act as an anti-recombinase to quell inappropriate recombinational repair during normal DNA metabolism.

Fig 1.

SDS/PAGE analysis showing induced expression of RecX in E. coli and at various stages during its purification. (A) Five micrograms of protein was separated by SDS/PAGE and visualized by staining with Coomassie blue. Lane M, molecular mass markers; lane 1, uninduced cell lysate; lane 2, induced cell lysate; lane 3, (NH₄)₂SO₄ pellet fraction; lane 4, purified RecX. (B) Western blot analysis. Lane 1, uninduced cell lysate; lane 2, induced cell lysate; lane 3, 5 μg of purified RecX; lane 4, 10 μg of purified RecX.

Fig 2.

RecX is associated with RecA in vivo. (A) RecA was immunoprecipitated (IP) with anti-RecX antibodies and immunoblotted (Blot) with anti-RecA antibodies. Lane 1, MtRecA. Immunoprecipitates from cell lysates of M. tuberculosis H37 Rv (lane 2) or M. smegmatis mc²155 (lane 5) with preimmune serum. The remaining lanes represent immunoprecipitates from cell lysates of indicated strains with anti-RecX antibodies. Lane 3, E. coli bearing M. tuberculosis recX; lane 4, M. tuberculosis H37 Rv; lane 6, M. smegmatis mc²155; lane 7, M. smegmatis mc²155 ΔrecA; lane 8, DNase I-treated cell lysate from M. smegmatis mc²155. (B) RecX was immunoprecipitated with anti-RecA antibodies and then immunoblotted with anti-RecX antibodies. Lane 1, purified RecX. Immunoprecipitates from cell lysates of M. tuberculosis H37 Rv (lane 2) or M. smegmatis mc²155 (lane 5) with preimmune serum. The remaining lanes contained immunoprecipitates of cell lysates of indicated strains using anti-RecA antibodies. Lane 3, E. coli bearing M. tuberculosis recX; lane 4, M. tuberculosis H37 Rv; lane 6, M. smegmatis mc²155; lane 7, M. smegmatis mc²155 ΔrecA; lane 8, DNase I-treated cell lysate from M. smegmatis mc²155.

Fig 3.

Interaction of RecX with RecA or SSB. RecX-Ni²⁺ agarose was incubated with the indicated RecA or SSB, and analyzed as described in Materials and Methods. Lanes marked FT, W1, and W2 correspond to flow-through (FT) and wash (fractions 1 and 2), respectively, from the column before elution. Lane M contained molecular mass markers, whose position (kDa) is indicated on the left. Under these conditions, SSB or RecA failed to interact with resin in the absence of tethered RecX.

Fig 4.

Effect of RecX on binding of RecA or SSB to DNA. (A) RecX failed to inhibit binding of MtRecA to ssDNA. Reactions were performed in the absence or presence of indicated amounts of RecX as described in Materials and Methods. (B) RecX failed to inhibit binding of MtSSB to ssDNA. Reaction was performed with 0.2 μM MtSSB and increasing concentrations of RecX, as described above. (C) RecX failed to bind single- or double-stranded DNA. Reaction mixtures (20 μl) containing10 μM circular single- or linear double-stranded DNA were incubated for 1 h at 37°C. Samples were analyzed by gel mobility shift assays, and visualized by staining with ethidium bromide. Lanes 1 and 10, DNA alone; lanes 2–9 and 11–18, DNA plus RecX at 0.25, 0.5, 0.75, 1, 1.5, 2, 3, and 4 μM, respectively. dsDNA, linear double-stranded DNA; lssDNA, linear single-stranded DNA

Fig 5.

RecX failed to inhibit binding of ATP by RecA. Reactions were performed and analyzed as described (4). Lane 1, RecA (0.75 μM); lane 2, MtRecX (1.5 μM); lane 3, MtSSB (1.5 μM). Lanes 4–7 contained 0.75 μM MtRecA and 0.75, 1, 1.2, and 1.75 μM RecX, respectively.

Fig 6.

RecX inhibits strand exchange promoted by RecA. (A) Effect of RecX on strand exchange promoted by EcRecA and EcSSB (lanes 3–8) or MtRecA and EcSSB (lanes 9–14). Lanes 3 and 9 represent controls in the absence of RecX. Lanes 4–8 represent reactions with EcRecA and EcSSB in the presence of 0.5, 1, 2.5, 5, or 7.5 μM RecX, and lanes 10–14 contained reactions with MtRecA and EcSSB in the presence of 0.025, 0.05, 0.075, 0.1, and 0.25 μM RecX, respectively. (B) Effect of RecX on strand exchange promoted by MtRecA and MtSSB (lanes 3–8) or MsRecA and MsSSB (lanes 9–14) in the absence (lanes 3 and 9) or presence (lanes 4–8 and 10–14) of 0.025, 0.05, 0.075, 0.1, and 0.25 μM RecX, respectively.

Fig 7.

RecX arrests strand exchange reaction in progress. Reactions were performed as described in the legend to Fig. 6, except that RecX (0.25 μM) was added to the reaction mixture at 0, 5, 10, 15, 30, and 45 min, respectively, after initiation of reaction by the addition of linear duplex DNA. All reactions were terminated after a 60-min incubation.