Repression of recA Induction by RecX Is Independent of the RecA Protein in Deinococcus radiodurans

RecX exhibited its regulation on recA transcription in the absence of RecA.

The recA-lacZ transcriptional fusion was constructed to analyze recA transcription in D. radiodurans. The recA promoter (about 203 bp) was obtained with the primers 5′-GTCTAGAGAAACACCAGCATGATCGC-3′ and 5′-GGGCCCACACTGATGATTTCTGCTA-3′, and cloned between the BglII and SpeI sites of pRADZ3 (13). The resulting plasmid, pAP1, was transformed into wild-type D. radiodurans R1, the recX mutant RecX⁻, the recA mutant RecA⁻, and the double mutant Rec(A+X)⁻ by a double crossover with selection for chloramphenicol (Chloromycetin) resistance (7). To avoid the differences in plasmid copy numbers, the transformation was carried out under the same conditions by electroporation, and the plasmid copy numbers in the four strains were quantified by reverse transcription-PCR (RT-PCR) and Southern blotting.

Expression of the lacZ reporter gene under the control of the recA promoter in E. coli or D. radiodurans was detected using X-Gal (5-bromo-4-chloro-3-inodyl-β-d-galactopyranoside; 40 μg/ml). β-Galactosidase activity in D. radiodurans was measured in toluene-permeabilized cells according to the method of Meima and Lidstrom (13). In the wild type, the lacZ reporter gene, driven by the recA promoter, retained basal level expression and could be induced by irradiation. In RecX⁻, the recA promoter enhanced its basal expression (P = 0.0023 versus recA basal expression in the wild type), and no induction was found after radiation treatment (Fig. 1A). This finding was in accord with a previous report that RecX inhibited the induction of the recA promoter (22). In RecA⁻, recA promoter transcription maintained its UV induction (P = 0.0021; n = 5), although its basal level expression increased slightly (P = 0.1057 versus recA basal expression in the wild type) (3, 21). RecX and RecA proteins exhibit different effects on the recA promoter. In the double mutant Rec(A+X)⁻, the recA promoter clearly upregulated its transcription (n = 5; P = 0.0027 versus recA basal expression in the wild type) and lost induction activity after UV irradiation. The absence of the RecA protein did not affect RecX regulation of promoter transcription. Thus, RecX appeared to regulate recA transcription independently of RecA.

• Open in new tab
• Download powerpoint

FIG. 1.

Transcription activity of the recA promoter. Plasmid pAP1 bearing a recA promoter-lacZ fusion (A) or plasmid pAP1-RecX bearing a recA promoter-lacZ fusion and an added recX gene (B) was transformed into the wild-type strain and three mutants. The promoter transcriptional level was assessed by β-galactosidase reporter activity. The plasmid copy numbers in the four strains were quantified by RT-PCR and Southern blotting. Lane 1, wild-type strain R1; lane 2, recX mutant RecX⁻; lane 3, recA mutant RecA⁻; lane 4, recA recX double mutant Rec(A+X)⁻. (C) Native recA transcript assay by RT-PCR. Strains were irradiated (gray) or not (white), and error bars represent mean ± standard deviation (SD) (n = 5).

We also investigated recA transcription in strains with an additional recX gene in the recA-lacZ fusion expression plasmid (22). With compensation or overexpression of the RecX protein, all four strains exhibited their induction activities (P < 0.05; n = 5), and their β-galactosidase activities were clearly inhibited (Fig. 1B). This result further indicated that the regulation of RecX on recA induction is independent of RecA activity.

Native recA transcript levels were also quantified by RT-PCR (Fig. 1C). Considering that the recA gene is disrupted by a Kam-resistant gene in the recA insertion mutant RecA⁻ and the recA recX double mutant Rec(A+X)⁻, we quantified the N terminus of recA (about 100 bp) as its native transcript in all four strains. Regardless of the presence or absence of RecA, the recA promoter upregulated the transcript level and lost its induction in the RecX mutant. This result was in agreement with the promoter-lacZ fusion analysis (Fig. 1A), indicating that repression of recA induction by RecX was independent of the RecA protein in Deinococcus radiodurans.

Without the N terminus, the RecX mutant protein (XL) maintained its interaction with RecA as well as its inhibition on the ATPase and DNA strand exchange activities of RecA and its inability to inhibit the transcription of the recA promoter. In addition to a conservative RecX superfamily domain in the C terminus, D. radiodurans RecX has an N-terminal domain (about 45 amino acids) that is not found in E. coli RecX. The N domain is rich in arginine, which endows RecX with a basic isoelectric point (pI) of about 9.96. Without its N terminus, the theoretical pI of RecX is about 5.89. To determine whether the N terminus of RecX was related to its transcription regulation activity, we cloned and expressed the D. radiodurans RecX protein that lacked the N terminus (named XL here) with the primers 5′-TTCATATGCCCCAAGACCCGCGAGGAGC-3′ and 5′-TTGGATCCTTACTCCTCGGCTTCGT-3′. Without the N domain, the size of the purified XL protein was about 20 kDa.

In most bacteria, including D. radiodurans, RecX can interact with RecA and inhibit RecA, in terms of its ATPase and DNA recombination ability. Here, RecA-XL interaction was determined by a His tag pulldown assay (His tag pulldown kit; Thermo Fisher Scientific) and the yeast two-hybrid assay (GAL4-based two-hybrid assay kit, Clontech, Japan). Assays were carried out according to each manufacturer's protocols. As shown in Fig. 2A, the XL protein could interact with the RecA protein and the absence of an N domain did not affect the binding of RecX and RecA. In addition to RecA-RecX interaction, the yeast two-hybrid assay also indicated interactions between RecX-RecX, RecX-XL, and XL-XL, suggesting that the RecX polymer formation was unrelated to its N domain.

• Open in new tab
• Download powerpoint

FIG. 2.

Biochemical characterization of XL protein. (A) Protein-protein interaction assay. His tag pulldown assay (left) and yeast two-hybrid assay (right) were carried out to inspect the interaction between RecA and XL proteins. Lane 1, His-tagged XL protein only; lane 2, His-tagged XL protein and non-His-tagged RecA; lane 3, His-tagged RecA protein and non-His-tagged XL; lane 4, His-tagged RecA protein only. (B) ATPase and DNA strand exchange assays of RecA in the absence or presence of gradient XL protein, with RecX as a control. jmDNA, joint-molecule DNA. (C) Compensation expression of RecX or XL protein in RecX⁻. Lane 1, wild-type strain R1; lane 2, recX mutant RecX⁻; lane 3, RecX⁻ with a compensation expression of RecX protein; lane 4, RecX⁻ with a compensation expression of XL protein. RecX or XL protein in the cell was quantified by Western blotting using RecX antibody. (D) DNA binding assay. DNA substrates and the amounts of protein are shown at the top of the images. The columns and error bars represent the mean ± standard deviation (SD) (n = 5).

The effects of XL on ATP hydrolysis promoted by RecA were investigated in a 20-μl reaction mixture (pH 7.0, 30°C, containing 50 mM HEPES buffer, 1 mM MgCl₂, 10 μM ATP, 0.2 μl [γ-³²P]ATP [10 mCi/ml], 40 fmol single-stranded DNA [ssDNA], 30 nM RecA protein, and 0 to 35 nM RecX [or XL] protein). After the reaction, the samples were cooled on ice and the reaction was stopped by adding 2 μl of 100 mM EDTA. ATP and P_i were separated by chromatography in 1 M formic acid-0.5 M LiCl. As seen in Fig. 2B, RecA ATPase activity was inhibited by the addition of XL protein to the reaction mixture, and the inhibition was similar to that seen with RecX. The RecA-dependent DNA strand exchange reaction was conducted as described previously (10). For the inhibition experiment, the indicated amounts of RecX (or XL) were added just before the addition of single-stranded DNA binding protein. Samples were electrophoresed in a 0.8% agarose gel and showed that XL clearly inhibited the DNA strand exchange activity (Fig. 2B) of RecA, similar to the effect of RecX. Based on these results, XL protein appeared to preserve the ability of RecX to inhibit RecA activity at the protein level, according to its sequence.

To study the effect of XL on recA transcription activity, the compensation expression of XL in the RecX⁻ mutant was studied. The XL DNA fragment was inserted into plasmid pAP1 between the sites NdeI and BamH, and the recombination plasmid pAP1-XL was used to detect the effect of XL on the recA transcription in a RecX⁻ mutant. Similarly, plasmid pAP1-RecX with a full recX gene was used as a control (Fig. 2C). In RecX⁻, the recA transcription clearly increased (lane 2), and the compensating expression of RecX protein repressed the transcription activity of recA (lane 3), but not that of XL (lane 4). XL protein lost its inhibition of the recA promoter, which suggested that the N-terminal domain of RecX might be involved in the regulation of recA expression.

Frequent recombination events initiated by RecA in stationary, growing bacteria under nonstressed conditions would be deleterious to the genomes. A tight negative regulatory control may be needed. In E. coli, LexA blocks recA expression on the transcriptional level, and RecX and DinI inhibit RecA activity on the protein level. A full understanding of the regulation of RecA in D. radiodurans remains elusive. In parallel to what has been described for E. coli, two lexA genes (dra0344 and dra0074) and a recX gene (dr1310) were found in D. radiodurans. The two D. radiodurans LexA proteins possess the conserved LexA domains and autocleavage activities promoted by RecA, but neither is involved in DNA damage-induction reactions (17, 23). D. radiodurans RecX protein exhibited the repression on the recA promoter besides its inhibition on RecA (22). So, an ideal model for the transcriptional regulation would be that the RecX protein could bind to a recA promoter and block recA transcription, similar to the function of LexA in E. coli. Gel shift assays (2) showed that purified D. radiodurans RecX could bind to ssDNA at a RecX concentration of ≥0.2 μM and to double-stranded DNA (dsDNA) at a RecX concentration of ≥0.4 μM, but not to the XL protein (Fig. 2D). The DNA binding ability of D. radiodurans RecX is related to its N terminus. Furthermore, recA promoter binding protein assays indicated that the recA promoter could pull down native RecX protein from the total protein extract of D. radiodurans. However, the DNA binding ability of the purified RecX protein was not exclusive toward the recA promoter in vitro (data not shown). A sequence assay showed that no conservative LexA box or other known DNA binding mode was found in the N terminus of RecX. Thus, we propose that RecX-specific binding of the recA promoter may require the assistance of another prosthetic group or proteins (not RecA).