Topoisomerase IV, not gyrase, decatenates products of site-specific recombination in Escherichia coli

Abstract

DNA replication and recombination generate intertwined DNA intermediates that must be decatenated for chromosome segregation to occur. We showed recently that topoisomerase IV (topo IV) is the only important decatenase of DNA replication intermediates in bacteria. Earlier results, however, indicated that DNA gyrase has the primary role in unlinking the catenated products of site-specific recombination. To address this discordance, we constructed a set of isogenic strains that enabled us to inhibit selectively with the quinolone norfloxacin topo IV, gyrase, both enzymes, or neither enzyme in vivo. We obtained identical results for the decatenation of the products of two different site-specific recombination enzymes, phage lambda integrase and transposon Tn3 resolvase. Norfloxacin blocked decatenation in wild-type strains, but had no effect in strains with drug-resistance mutations in both gyrase and topo IV. When topo IV alone was inhibited, decatenation was almost completely blocked. If gyrase alone were inhibited, most of the catenanes were unlinked. We showed that topo IV is the primary decatenase in vivo and that this function is dependent on the level of DNA supercoiling. We conclude that the role of gyrase in decatenation is to introduce negative supercoils into DNA, which makes better substrates for topo IV. We also discovered that topo IV has an unexpectedly strong DNA relaxation activity that, together with gyrase and topo I, is able to set the supercoiling levels in Escherichia coli.

Figure 1

Experimental system for measuring the unlinking of catenanes produced by λ Int recombination. Cells were grown at 30°C to a density of 70 Klett units and shifted to 43°C where Int is expressed but is inactive. Upon downshift to 30°C, Int recombines the substrate plasmid pJB3.5d to make multiply interlinked catenanes. A type-2 topoisomerase unlinks the catenanes to release free 0.6- and 2.9-kb circles.

Figure 2

Effect of norfloxacin on Int recombination. The experiment was as described in the legend of Fig. 1. Plasmid DNA was isolated, subjected to electrophoresis on high resolution agarose gels, Southern blotted, and quantified by PhosphorImager analysis. The total amount of plasmid DNA label was normalized to 100%. The percentage of the total recombined plasmid is shown for various time points and drug concentrations [Nor] in each of the six strains. Time after shift down to 30°C is represented as black wedges. Samples were analyzed at the following times after the addition of drug: 0 min (▪), 20 min (squares in box), 30 min (white box), 45 (hatched box), and 60 min (dotted). The graphs are presented in a matrix array depending upon the states of the topo IV and gyrase alleles. (parC⁺ gyrA⁺) LZ36; (parC⁺ gyrA^L83) LZ35; (parC^L80 gyrA⁺) LZ34; (parC^L80 gyrA^L83) LZ33; (parC^K84 gyrA⁺) LZ38; (parC^K84 gyrA^L83) LZ37. (N.D.) Not determined. Parts of these experiments were repeated three times. The exact values varied somewhat, but the relative values were always the same. For example, the 30 μm drug, 20 min point for the double wild-type strain was done three times. The amount of recombination averaged 30% ± 7% and the amount of catenation averaged 75% and varied <4%. The results here and in Figs. 3 and 5 are from one extensive experiment performed all on the same day.

Figure 3

Effect of norfloxacin on plasmid DNA supercoiling levels. Supercoiled plasmid DNA from the experiment described in Fig. 2 was separated and visualized on agarose gels containing 0, 2, 4, or 8 μg/ml chloroquine and compared with plasmids of known σ. Shown are the σ values as quantified by the band counting method (Keller 1975). A ς of zero was taken as the midpoint of the topoisomer distribution after plasmid DNA relaxation in vitro with calf thymus topo I. The time points and symbols are the same as in Fig. 2.

Figure 4

Effect of DNA supercoiling (σ) on Int recombination. The amount of recombination as a function of σ was assessed by plotting all the data from Figs. 2 and 3. The line is drawn for ease of visualization.

Figure 5

Effect of norfloxacin on decatenation of Int recombination products. The amount of recombined DNA from Fig. 2 was further broken down into catenanes and free circles (decatenated). The percentage of the total recombined plasmid that was catenated is shown.

Figure 6

Effect of DNA supercoiling (σ) on decatenation. The correlation between decatenation and σ is shown for every parC⁺ data point in Fig. 5.

Figure 7

Kinetics of recombination and decatenation. (A) Rate of recombination in top10 parC⁺ gyrA⁺ (LZ53; •) and top10 parC⁺ gyrA^L83 (LZ54; ○) strains. The experiment was as outlined in Figs. 1 and 2. The drug concentration was 90 μm. (B) Amount of catenation in top10 parC⁺ gyrA⁺ (•) and top10 parC⁺ gyrA^L83 (○) strains. Catenane turnover in a parC^K84 gyrA^L83 strain (LZ37; ▪) is shown for comparison. From the relative slopes, catenane unlinking when gyrase is active and topo IV is inhibited (○) is 3.5-fold faster than when both enzymes are inhibited (•) and 25-fold faster than when topo IV is uninhibited (▪).

Figure 8

Experimental system for measuring decatenation of the products of resolvase recombination. The cells were grown at 30°C and then shifted to 43°C to induce resolvase expression. The enzyme recombines the two directly repeated sites on the substrate plasmid DNA and the product is a singly linked (two-noded) catenane. The sizes of the rings were 2.75 and 3.2 kb. A type-2 topoisomerase unlinks the catenanes to generate free circles. Drug (norfloxacin) was added after recombination but before most unlinking.

Figure 9

Effect of topo IV on decatenation of resolvase recombination products. The fraction of plasmid DNA that was catenated after 10 min at 43°C is shown as a function of norfloxacin (Nor) concentration. Two strains were assayed: parC⁺ gyrA^L83 (LZ29; •) and parC^L80 gyrA^L83 (LZ27; ▪).