Contribution of the ATP Binding Site of ParE to Susceptibility to Novobiocin and Quinolones in Streptococcus pneumoniae

Abstract

In Streptococcus pneumoniae, an H103Y substitution in the ATP binding site of the ParE subunit of topoisomerase IV was shown to confer quinolone resistance and hypersensitivity to novobiocin when associated with an S84F change in the A subunit of DNA gyrase. We reconstituted in vitro the wild-type topoisomerase IV and its ParE mutant. The ParE mutant enzyme showed a decreased activity for decatenation at subsaturating ATP levels and was more sensitive to inhibition by novobiocin but was as sensitive to quinolones. These results show that the ParE alteration H103Y alone is not responsible for quinolone resistance and agree with the assumption that it facilitates the open conformation of the ATP binding site that would lead to novobiocin hypersensitivity and to a higher requirement of ATP.

Topoisomerase IV and DNA gyrase are bacterial type II topoisomerases that modify DNA topology during the replication process by unlinking DNA and facilitating chromosome segregation (33). DNA gyrase, encoded by gyrA and gyrB, forms an A₂B₂ tetramer that removes the positive supercoils introduced during DNA replication and helps to maintain the steady-state balance of negative supercoiling of the bacterial DNA (30). Topoisomerase IV, encoded by parC and parE, forms a C₂E₂ tetramer. Its primary role is the decatenation of daughter chromosomes following DNA replication, and other functions are DNA relaxation and double-strand DNA breakage (1, 6, 14, 33). Both enzymes are the target of two antibiotic families: the quinolones, such as ciprofloxacin, sparfloxacin, and grepafloxacin, and the coumarins, such as novobiocin. The quinolones form a ternary complex with topoisomerases in the presence of DNA, resulting in double-stranded breaks (4). The coumarins inhibit ATP-dependent functions of these enzymes by preventing binding and hydrolysis of ATP (27).

In Streptococcus pneumoniae, it has been shown that the primary and the secondary targets of quinolones are either the gyrase or the topoisomerase IV, depending upon the molecule (21, 29). Mutants resistant to quinolones most frequently harbor mutations in the quinolone resistance-determining regions (QRDR) of GyrA and ParC and less frequently in that of ParE (7). The level of quinolone resistance increases when a mutation is present in the secondary target in addition to one in the primary target (21, 29). In S. pneumoniae, in vitro-selected resistance to novobiocin is associated with the S127L substitution in the GyrB subunit (17). This position is located upstream from and close to positions 136 and 164 described in the ATP binding domain of GyrB in Escherichia coli and also involved in coumarin resistance (3). The substitution in ParE of E. coli at a position equivalent to that of position 136 in GyrB resulted in a decreased inhibition of topoisomerase IV by novobiocin (6).

In previous studies (11, 26), we showed in S. pneumoniae that the H103Y substitution in the N terminus of the ParE subunit of the topoisomerase IV resulted in an original phenotype of resistance only if it was associated with an S84F substitution in GyrA. This phenotype was associated with no change in ciprofloxacin susceptibility, decreased susceptibility to sparfloxacin and grepafloxacin, and increased susceptibility to novobiocin (11, 26). This position is located in the 43-kDa ATP binding domain (26), which is homologous to the ATPase catalytic site of the E. coli GyrB subunit of DNA gyrase (2).

The crystal structures of the 43-kDa N-terminal GyrB domain of E. coli in complex with ADPNP and of Thermus thermophilus in complex with novobiocin have shown that the ATP binding site adopts either a closed or open conformation (2, 13). Based on these structures, the modeling of the 43-kDa domain of ParE of S. pneumoniae suggested that the substitutions H103Y and H103F would affect preferentially the ATP-bound closed conformation of the ATP active site (26).

In order to better understand the effect of the H103Y mutation, we purified the topoisomerase IV subunits of the wild type (WT) and the H103Y ParE mutant of S. pneumoniae and compared the different catalytic activities of the reconstituted topoisomerases as well as their inhibition by quinolones and novobiocin.

Purification of the wild-type ParE and its H103Y mutant.

S. pneumoniae R6 (WT) is a susceptible derivative of the nonencapsulated Rockefeller University strain R36A (28); R6^{GyrA S84F} harbors the S84F substitution in GyrA, and R6^{GyrA S84F/ParE H103Y} harbors, in addition, the H103Y substitution in ParE (11). Genomic DNA (10) of S. pneumoniae R6 and R6^{GyrA S84F/ParE H103Y} was used as template to clone the WT parC and parE genes and the H103Y parE mutant allele, respectively. parC and parE genes were inserted into the pTYB-1 fusion protein expression vector (New England BioLabs Ltd.) that harbors an intein self-cleavable tag. The parC (19) gene was amplified by using the forward primer PDParC1, 5′-AAAAAAAACATATGTCTAACATTCAAAACATGTCCCT-3′ (NdeIsite underlined), and the reverse primer PDParC3, 5′-TATTATTGCTCTTCCGCATTTATCTTCAGTAACTACTTCCTG-3′ (SapI site underlined). The parE alleles (19) were amplified by using the forward primer PDParE1, 5′-AAAAACATATGTCAAAAAAGGAAATC-3′ (NdeI site [underlined], converting the GTG initiation codon to ATG), and the reverse primer PDParE6, 5′-GGTGGTTGCTCTTCCGCAACCAAACACTGTCGCTTCTTCTAG-3′ (SapI site underlined), which corresponds to the sequence downstream of the stop codon except that an extra C-terminal glycine was added (italicized) for efficient self-cleavage of intein. The amplified WT parC and the parE and H103Y mutant parE products were digested and ligated into the NdeI and SapI sites of pTYB-1. The resulting plasmids were introduced into E. coli ER 2566 (New England BioLabs Ltd.) to express the ParE or ParC fusion proteins. Cells were grown to an optical density at 650 nm of 0.6 to 0.8 and then overnight at 16°C after induction with 0.5 mM isopropyl-β-d-thiogalactopyranoside (Sigma). After centrifugation, the pellet was suspended in 30 ml of buffer A (20 mM Tris-HCl [pH 7.5], 500 mM NaCl, 0.5 M EDTA [pH 8.0]). After disruption of the cells and centrifugation at 20,000 × g for 60 min at 4°C, the supernatant containing the intein-ParE or -ParC fusion was loaded onto a column of chitin beads (New England BioLabs). On-column cleavage was conducted in the presence of cleavage buffer (buffer A and 50 mM dithiothreitol) during 16 h at 4°C. Elution resulted in ParC and ParE purified to apparent homogeneity (>97%) (data not shown). To reconstitute the topoisomerase IVs, the WT ParC subunit was combined with either the WT ParE or the H103Y mutant ParE. Only the combination of ParC and ParE subunits yielded topoisomerase IV activities, i.e., ATP-dependent decatenation, ATP-dependent relaxation, and ATP-independent double-stranded DNA cleavage.

Effect of the ParE H103Y substitution on decatenation and relaxation of the reconstituted topoisomerase IV.

Catalytic activities of topoisomerase IV were tested as described previously (20, 21). For the decatenation assay, 0.4 μg of kinetoplast DNA (kDNA; TOPOGEN, Inc., Columbus, Ohio) and 2 U of topoisomerase IV were incubated at 37°C for 1 h. For the relaxation assay, 0.4 μg of supercoiled pBR322 (Roche-Boehringer, Mannheim, Germany) and 2 U of topoisomerase IV were incubated at 37°C for 1 h.

The specific activities of WT ParE and mutant H103Y ParE were first determined using an excess of ParC and an excess of ATP (≥1 mM). Under these conditions the WT and H103Y mutant ParE had identical specific activities either in the decatenation assay (3 × 10⁵ U/mg) or in the relaxation assay (4 × 10⁴ U/mg). These activities were within the range reported for the WT topoisomerase IV of S. pneumoniae (5, 20) and, as described previously, the relaxation activity was even about 11-fold lower than the decatenation activity for the ParE mutant topoisomerase IV (8, 11, 20). Since the specific topoisomerase IV activities of the homologs containing WT ParE and H103Y mutant ParE were determined in the presence of an excess of ATP, this could have hidden more-subtle alterations in ATP binding. We subsequently tested kDNA decatenation in an ATP-requiring assay by using a range of low ATP concentrations (Fig. 1; Table 1). At 40 μM ATP, which is the optimal concentration for E. coli topoisomerase IV-mediated decatenation (23), 10% of the kDNA was decatenated in 1 h by the WT topoisomerase IV of S. pneumoniae, but the H103Y ParE mutant topoisomerase IV did not display any decatenation activity (Fig. 1). These results demonstrate an ATP requirement at subsaturating ATP levels that has been previously described in altered eucaryotic type II topoisomerases, although the mutations were located at positions different from H103Y in ParE (22). A fourfold decrease in topoisomerase activity was also previously observed in E. coli GyrB mutants (6). Our results agree with the suggestion, based on the modeling of the 43-kDa N-terminal ParE domain of S. pneumoniae (26), that the H103Y substitution disturbs the ATP-bound conformation of the ATPase active site. In contrast, there was no difference in the ATP requirement when the WT and the altered topoisomerases were tested for relaxation activity using a wide range of concentrations (40 μM to 1.5 mM) of ATP. For both enzymes no activity was observed in the absence of ATP, and 80 μM ATP was necessary to obtain 50% relaxation (Table 1). This could reflect a differential involvement of the ATPase active site in the DNA-topoisomerase IV complex in the two distinct functions of relaxation and decatenation. These two functions are different in that intramolecular cleavage is involved in relaxation and intermolecular cleavage is involved in decatenation (15). The trapping of the T segment DNA should be different in each function, and the conformation of the N terminal of ParE might be involved in this trapping (25). Since the presence of the H103Y mutation in ParE would favor an open conformation of the ATPase active site (26), the positioning of the T segment should not be optimal. It would result, at subsaturating concentrations of ATP, i.e., when the site cannot be completely closed, in a decrease in the decatenation activity of topoisomerase IV.

FIG. 1.

ATP requirement for topoisomerase IV-catalyzed decatenation. Reactions were done in the presence of 2 U of WT topoisomerase IV or mutant topoisomerase IV (H103Y) and either 0, 40, 80, 120, 160, 320, 640, or 1,000 μM ATP. N, kinetoplast network DNA; M, released relaxed minicircles.

TABLE 1.

Susceptibility to fluoroquinolones and novobiocin of S. pneumoniae WT and mutants and activities of the respective topoisomerase IV

Agent and assay	Concn for strain (topoisomerase IV form)
	R6a (WT)	R6GyrA S84F (WT)b	R6GyrA S84F/ParE H103Y (mutant)c
Sparfloxacin
MIC (μg/ml)	0.25	1	4
IC50 (μg/ml)
Decatenation	4	—g	4
Relaxation	NDf	—	ND
CC25d (μg/ml)	0.25	—	0.25
Grepafloxacin
MIC (μg/ml)	0.25	1	4
IC50 (μg/ml)
Decatenation	4	—	4
Relaxation	16-32	—	16-32
CC25 (μg/ml)	ND	—	ND
Novobiocin
MIC (μg/ml)	0.25	0.25	0.125
IC50 (μg/ml)
Decatenation	24	—	12
Relaxation	32	—	32
ATP (μM) requiremente
Decatenation	160	—	320
Relaxation	80	—	80

^aWild-type S. pneumoniae strain used to reconstitute the WT topoisomerase IV.

^bMutant derived from the WT strain R6 and harboring the mutation S84F in the QRDR of GyrA with a WT topoisomerase IV.

^cMutant derived from the WT strain R6 and harboring the mutation S84F in the QRDR of GyrA and the H103Y mutation in ParE. Its ParE product was used with the WT ParC to reconstitute the mutated topoisomerase IV.

^dConcentration of fluoroquinolone that converts 25% of 0.4 μg of pBR322 DNA to the linear form.

^eConcentration of ATP required to decatenate 50% of 0.4 μg of kDNA or relax 50% of 0.4 μg of supercoiled pBR322 in 1 h at 37°C.

^fND, not determined.

^gNot relevant, since the topoisomerase IV is the same as that of the WT R6.

Topoisomerase IV containing H103Y ParE can bind ATP and close its ATP-dependent clamp.

ATP binding is required to close the ATP-dependent clamp of the type II topoisomerases (12). Since the presence of the H103Y substitution in ParE could displace the equilibrium in favor of the open conformation (26), it could result in a defect in dimerization of the ParE subunits. To examine this hypothesis, we performed a protein-protein cross-linking assay exactly as described by Nurse et al. (18) to assess the ability of the ParE subunits to dimerize in the presence of ADPNP, a nonhydrolyzable ATP analog. Accordingly, the ParE proteins (WT and H103Y mutant) were treated with dimethylsuberimidate di-HCl (DMS; Sigma) in the presence or absence of ADPNP (Sigma). Only in the presence of ADPNP did both the WT and the H103Y ParE proteins produce the same apparent quantity of a complex of ca. 230-kDa protein (Fig. 2). Identical results were obtained using different times of incubation for dimerization and over a wide range of ADPNP and DMS concentrations (data not shown). Since a ParE dimer should have a mass of 140.4 kDa, it is conceivable, according to Nurse et al. (18), that this complex was a trimer or resulted from a particular migration of a dimer. These results indicate that the ParE H103Y mutant subunit of topoisomerase IV binds ADPNP and ATP and closes its ATP-dependent active site in a manner similar to that of the WT ParE.

FIG. 2.

Dimerization of ParE WT and ParE H103Y in the presence of ADPNP. The indicated ParE protein (WT or H103Y) at 3 μM was treated with DMS (7 mg/ml) in either the presence (+) or absence (−) of 1 mM ADPNP. The products were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis through a 9% gel with Coomassie blue staining. The open arrowhead indicates the migration position of the ParE monomer. The solid arrowhead aligns the migration position of the putative ParE cross-linked dimer.

Effects of antibiotics on the activities of the WT and H103Y ParE-containing topoisomerase IV.

After recombination in the WT S. pneumoniae R6, we had demonstrated that the H103Y substitution in ParE was responsible for a particular phenotype only in the presence of the S84F substitution in GyrA, i.e., an increased susceptibility to novobiocin and a further decreased susceptibility to sparfloxacin and grepafloxacin (Table 1) (11, 26). In order to evaluate the role of the H103Y substitution at the biochemical level, we compared the inhibitory effect of sparfloxacin, grepafloxacin, and novobiocin on the decatenation and relaxation activities of topoisomerase IV as well as the stimulating effect of these quinolones on DNA breakage.

Increases in the concentration of the quinolones resulted in a dose-dependent inhibition of decatenation and relaxation. The concentrations of sparfloxacin and grepafloxacin required to inhibit 50% (IC₅₀) of decatenation were identical for the WT and the mutant topoisomerase IV (Table 1). The IC₅₀s of grepafloxacin for relaxation were also identical for the WT and H103Y mutant topoisomerase IV. Since DNA cleavage is the relevant quinolone-mediated cytotoxic lesion (14, 32) which, for S. pneumoniae, has been shown to occur earlier than the inhibition of the other topoisomerase IV activities (20, 21), we examined the quinolone-induced DNA cleavage arising from the stabilization of the cleavable complex. DNA cleavage assays were the same as the relaxation assays, except that ATP was omitted. In each reaction, the WT ParC (0.5 μg) and the WT or the mutant topoisomerase IV ParE (1 μg) were incubated with supercoiled pBR322 (0.4 μg) for 1 h at 25°C in the absence or presence of increasing concentrations of sparfloxacin to promote topoisomerase IV-mediated DNA breakage. The concentrations of sparfloxacin required for the conversion of 25% of the pBR322 DNA to the linear form (CC₂₅ in Table 1) were similar for the mutant and the WT topoisomerase IV, and these values were in accordance with those of previous studies that used a WT topoisomerase IV (20, 21).

Genetic studies of S. pneumoniae have shown that the mutations in the QRDR of either GyrA, ParC, or ParE, alone, but not of GyrB, result in a decreased susceptibility depending upon the fluoroquinolone tested (7, 24, 29, 31). Purified topoisomerases of S. pneumoniae containing altered GyrA or ParC subunits showed significantly decreased inhibition by quinolones for supercoiling and decatenation activities as well as reduced DNA breakage (32, 21). So far, no study has explored the quinolone inhibitory activity on topoisomerase IV containing a mutationally altered ParE. The particular H103Y substitution studied in this work is located outside of the QRDR and in the ATPase domain of ParE. Fluoroquinolone resistance associated with the H103Y substitution in ParE was expressed only when the parE mutation was associated with a mutation in the QRDR of GyrA (26). This is not totally surprising, since gyrase is the primary target of sparfloxacin and grepafloxacin, while topoisomerase IV is only the secondary target (16, 29). Nonetheless, there was no decrease in the inhibitory effect of these quinolones on the catalytic activities of the topoisomerase IV reconstituted with the mutated H103Y subunit. One possibility could be that the mutant topoisomerase IV, which is less active in the presence of a suboptimal concentration of ATP, would form less-stable ternary complexes. This would lead to fewer lethal double-strand breaks and therefore to a decreased activity of fluoroquinolones (9).

No difference was observed between the WT and the mutant topoisomerase IV in the ATP-dependent relaxation assay when the inhibition by novobiocin was tested (Table 1). However, in the decatenation assay, the mutant topoisomerase IV was more sensitive to the inhibition by novobiocin than the WT topoisomerase IV was (Table 1; Fig. 3). Knowing that decatenation is the essential activity of topoisomerase IV in vivo (14), this biochemical result correlates with the MIC results that showed that the H103Y substitution in ParE was associated with a twofold increase in the susceptibility to novobiocin (Table 1) (11, 26). According to the model of Sifaoui et al. (26), a mutation at position 103 results in a preferential open conformation of the loop closing the ATP active site of ParE. Since the ATP and novobiocin binding sites overlap (6), this conformation could entail different effects: a higher need of ATP for decatenation of the mutated topoisomerase IV as well as a facilitated access of novobiocin to its ParE subunit. This would then explain the better inhibition of decatenation by novobiocin, a competitive inhibitor of ATP (6), and thereby the increase in susceptibility to this antibiotic in vivo.

FIG. 3.

Inhibition by novobiocin of topoisomerase IV-catalyzed decatenation. Reactions were done in the presence of 2 U of topoisomerase IV WT or mutated topoisomerase IV (H103Y) and either 0, 8, 16, 24, 32, or 40 μg of novobiocin/ml. N, kinetoplast network DNA; M, released relaxed minicircles.