ABSTRACT
Streptococcus pneumoniae gyrA and gyrBgenes specifying the DNA gyrase subunits have been cloned into pET plasmid vectors under the control of an inducible T7 promoter and have been separately expressed in Escherichia coli. Soluble 97-kDa GyrA and 72-kDa GyrB proteins bearing polyhistidine tags at their respective C-terminal and N-terminal ends were purified to apparent homogeneity by one-step nickel chelate column chromatography and were free of host E. coli topoisomerase activity. Equimolar amounts of the gyrase subunits reconstituted ATP-dependent DNA supercoiling with comparable activity to gyrase of E. coli and Staphylococcus aureus. In parallel, S. pneumoniae topoisomerase IV ParC and ParE subunits were similarly expressed in E. coli, purified to near homogeneity as 93- and 73-kDa proteins, and shown to generate efficient ATP-dependent DNA relaxation and DNA decatenation activities. Using the purified enzymes, we examined the inhibitory effects of three paradigm fluoroquinolones—ciprofloxacin, sparfloxacin, and clinafloxacin—which previous genetic studies with S. pneumoniae suggested act preferentially through topoisomerase IV, through gyrase, and through both enzymes, respectively. Surprisingly, all three quinolones were more active in inhibiting purified topoisomerase IV than gyrase, with clinafloxacin showing the greatest inhibitory potency. Moreover, the tested agents were at least 25-fold more effective in stabilizing a cleavable complex (the relevant cytotoxic lesion) with topoisomerase IV than with gyrase, with clinafloxacin some 10- to 32-fold more potent against either enzyme, in line with its superior activity againstS. pneumoniae. The uniform target preference of the three fluoroquinolones for topoisomerase IV in vitro is in apparent contrast to the genetic data. We interpret these results in terms of a model for bacterial killing by quinolones in which cellular factors can modulate the effects of target affinity to determine the cytotoxic pathway.
The recent development of new fluoroquinolones effective against Streptococcus pneumoniaeis a potentially important advance in the management of pneumococcal disease (3, 20, 34). Several of these agents are more active than ciprofloxacin and have either been approved or are in the late stages of clinical trials, including levofloxacin, sparfloxacin, trovafloxacin, grepafloxacin, and clinafloxacin. Given that fluoroquinolones act by blocking DNA synthesis, they should be effective against both penicillin-sensitive and -resistant S. pneumoniae (13, 28, 40). Progress in this area has focused attention on the nature and cellular consequences of quinolone interactions with their pneumococcal targets, the essential type II topoisomerases, DNA gyrase and DNA topoisomerase IV (41).
Previous work in Escherichia coli has established that both gyrase and topoisomerase IV act by a double-strand DNA break and play important roles in facilitating DNA transactions, especially DNA replication (1, 6, 10, 17, 21, 44). Gyrase, a complex of two GyrA and two GyrB subunits encoded by the gyrA andgyrB genes, introduces negative supercoils into DNA in a reaction driven by ATP hydrolysis. The enzyme is essential for initiation of DNA replication and plays a key role in elongation, presumably by removing positive supercoils arising from DNA unwinding at the replication fork. Though gyrase was formerly implicated in the unlinking of daughter chromosomes, it now appears that this function is largely performed by the ATP-requiring enzyme topoisomerase IV, a C2E2 tetramer encoded by the parCand parE genes (1, 44). Quinolones inhibit both enzymes by forming a ternary complex of drug, enzyme, and DNA—“the cleavable complex”— that, on treatment with detergent and proteinase K, generates double-strand DNA breaks (6, 9). Similarly, it is thought that cellular processes acting on the ternary complex in vivo result in the formation of an irreparable double-stranded DNA break, thereby triggering bacterial cell death (18, 19).
Resistance to quinolones in E. coli usually involves point mutations in defined regions of the GyrA or GyrB proteins, termed the quinolone resistance-determining regions, or QRDRs (5, 26, 42,43). Mutations in the equivalent regions of the ParC or ParE proteins occur subsequent to those in gyrase and are associated with very high level resistance to the drugs (12). These findings originally led to the conclusion that gyrase is invariably the primary target of the quinolones. However, studies with Staphylococcus aureus and S. pneumoniae have discounted this idea (7, 8, 27, 29-32). Unlike E. coli, in which all quinolones tested thus far target DNA gyrase, it appears that, inS. pneumoniae, bacterial killing can proceed through either gyrase, topoisomerase IV, or both, depending on the structure of the quinolone (31, 32). This important conclusion has come primarily from the characterization of stepwise-selected quinolone-resistant S. pneumoniae mutants. Thus, on challenge of S. pneumoniae with ciprofloxacin,parC or parE mutations appear before those ingyrA, suggesting the drug acts preferentially through topoisomerase IV (11, 16, 25, 29, 30, 33, 39). In contrast, challenge with sparfloxacin selects for gyrA mutations before those in parC, indicating a primary role of drug interactions with gyrase (31). For clinafloxacin, killing proceeds through gyrase and topoisomerase IV, with a modest preference for gyrase (32). The molecular basis underlying these different drug specificities is currently unknown, but it presumably reflects differential enzyme-drug interactions and/or more complex factors, such as differential lethalities of enzyme inhibition.
In an effort to understand the different in vivo mechanisms of ciprofloxacin, sparfloxacin, and clinafloxacin, we have examined the contribution of enzyme inhibition to drug action. Here, we report the overexpression and purification of S. pneumoniae GyrA, GyrB, ParC, and ParE subunits; reconstitution of highly active gyrase and topoisomerase IV proteins; and their differential responses to quinolones.
MATERIALS AND METHODS
Bacterial strains, plasmids, and DNA substrates. S. pneumoniae 7785 and the conditions for its growth have been described previously (30). E. coli XL-Blue and plasmid pBluescript were used to construct libraries and to subclone DNA inserts. Plasmid pCRII (Invitrogen) was used to clone inverse PCR (IPCR) products in E. coli XL-Blue. Vectors pET-19b and pET-29a (Novagen) were used to construct plasmids for overexpression ofS. pneumoniae GyrB and GyrA and ParC and ParE proteins inE. coli host BL21(λDE3)pLysS. Some intermediate plasmid constructs were propagated in E. coli DH5α. Supercoiled pBR322 was prepared as described previously (30), and relaxed-DNA pBR322 was obtained by incubation with calf thymus topoisomerase I (Life Technologies). Kinetoplast DNA (kDNA) fromCrithidia fasciculata was purchased from TOPOGEN, Inc., Columbus, Ohio.
Chemicals and reagents.Ciprofloxacin-HCl and clinafloxacin-HCl were kindly provided by Bayer UK, Newbury, United Kingdom, and by Parke-Davis Co., Ann Arbor, Mich. Sparfloxacin was a kind gift from Dainippon Pharmaceutical Co., Suita, Japan. Oligonucleotide primers were synthesized by Oswel, Ltd., University of Southampton, Southampton, United Kingdom.
Molecular cloning of the S. pneumoniae 7785gyrA gene.In previous work, we have described the isolation and characterization of the parE, parC, and gyrB genes from S. pneumoniae 7785 (30). To obtain the gyrA gene from this strain, a radiolabeled 382-bp gyrA PCR product amplified from S. pneumoniae 7785 (29) was used to probe a Southern blot of 7785 genomic DNA and was found to hybridize to 4.0-kbPstI and 2.5-kb HindIII fragments (not shown). Size-selected S. pneumoniae 7785HindIII and PstI libraries in pBluescript SK were constructed in E. coli XL1-Blue and screened with thegyrA PCR product. Due to plasmid instability, only the 2.5-kb HindIII fragment could be isolated intact in plasmid pXP11. DNA sequence analysis of the pXP11 insert showed it contained two incomplete open reading frames (ORFs), one of which encoded the N-terminal 411 residues of GyrA protein. To obtain the 3′ end of the gyrA gene, IPCR was used, employing forward primer VGA14 (5′ TGAAACGGATGCGGAAGCTCAAGC [gyrAnucleotide positions 1190 to 1213]) and reverse primer VGA13 (5′-ACTCTGACTGTGCAGACGCTGACC [−406 to −429 adjacent to aPstI site]). S. pneumoniae 7785 genomic DNA was digested to completion with PstI, and the restriction fragments were circularized by ligation (30). The fragments were then used as a template in IPCR with primers, Taq DNA polymerase, and 1.5 mM MgCl2 (30). The conditions were 94°C for 1 min, 45°C for 1 min, and 74°C for 2 min (30 cycles). A 2.4-kb IPCR product was obtained, cloned directly into plasmid pCRII, and transformed into E. coli XL-Blue. Plasmid pXP12 was recovered from one of the ampicillin-resistant colonies.
The inserts in pXP11 and pXP12 specified a 4.7-kb region of theS. pneumoniae chromosome whose DNA sequence was determined, and two divergent ORFs were identified. An incomplete ORF, specified by the 5′ end of the pXP11 insert, encoded a 325-residue protein which was identical to the l-(+)-lactate dehydrogenase of S. pneumoniae, a 328-residue enzyme that catalyzes the fructose-1,6-diphosphate-dependent interconversion of pyruvate and lactate (15). The second ORF encoded the 822-residue GyrA protein, which exhibits a predicted molecular mass of 92 kDa. Except for seven amino acid differences most likely arising from strain polymorphisms, the strain 7785 GyrA sequence is identical to that recently reported by Balas et al. (2): Ile-489, Lys-537, Lys-642, and Thr-653 and the three consecutive conserved residues Gly-618–Ile–Val in our sequence are replaced in that of Balas et al. by Val, Glu, Gln, Ala and Val-Leu-Leu, respectively.
DNA sequence analysis.Cloned S. pneumoniae DNA fragments were sequenced on both strands by the chain termination method (36). The plasmid pXP11 insert was sequenced by using T7 and internal primers. Sequence at the extreme 3′ end of the pXP12 insert was obtained by using T7 and SP6 primers. This information was then used to design primers for asymmetric PCR (AsPCR) by using the proofreading Vent DNA polymerase and strain 7785 genomic DNA as a template. AsPCR products were then sequenced directly (31).
Construction of GyrA- and GyrB-expressing plasmids.PCR was used to amplify the S. pneumoniae gyrA and gyrBgenes for insertion into expression vectors pET-29a and pET-19b.NdeI sites (CA′TATG) were engineered into each of the forward primers overlapping the ATG initiation codon ofgyrA and gyrB. The gyrB gene was amplified by using forward primer (P7164) 5′-AGAAAAAGGAATCATATGACAGA AG (NdeI site underlined) and reverse primer (P7165) 5′-AGGGAACTACTTCTCGAGATTTTTTA (XhoI site underlined). The gyrA gene was assembled from two PCR products obtained as follows. The 5′ end of gyrA was amplified with forward primer VGA35 (5′-ATGAGGCATTTACATATGCAGGATAAAAATTTAGTG) and reverse primer VGA22 (5′-AGCCCTTTGGCAGTCCGACC [nucleotide positions 1722 to 1741; i.e., 3′ to an XhoI site ingyrA at nucleotide 1473]). The 3′ end of gyrAwas obtained by using forward primer VGA17 (5′-ACAGAGTTGATGGTTGGAC [nucleotide positions 1442 to 1460]) and reverse primer VGA36 (5′-GAGACACTCGAGTTCACCTTCTGTTTCGTTTTC [XhoI site underlined]). PCR was carried out with genomic DNA from strain 7785 by using Vent DNA polymerase in the presence of primers and 1.5 mM MgCl2. The PCR conditions were denaturation at 94°C for 1 min, annealing for 1 min at 45°C (for gyrA primers) or 52°C (for gyrB primers), and polymerization at 72°C for 3 min. Reactions were performed over 30 cycles. The gyrB and 5′ gyrA PCR products were digested with NdeI and XhoI, purified by electrophoresis in low-gelling temperature agarose, and recovered. The resulting gyrB and 1,473-bp gyrA products were ligated into NdeI-XhoI-cut pET-19b and pET-29a and transformed into E. coli DH5α, and resistant colonies were selected on plates containing ampicillin or kanamycin, respectively. This procedure allowed recovery of gyrBexpression plasmid pXP9 and pXP55 bearing a partial gyrAgene. The 3′ gyrA PCR product was digested withXhoI, and the resulting 990-bp fragment was ligated intoXhoI-linearized pXP55, whose 5′ ends had been dephosphorylated by using calf intestinal alkaline phosphatase. After transformation of E. coli DH5α, recovery of plasmids from kanamycin-resistant colonies followed by restriction analysis identified expression plasmid pXP10, in which the gyrA gene had been correctly assembled. The translation initiation and termination regions of the recombinant plasmids pXP9 and pXP10 were sequenced to confirm that fragments had been inserted in frame.
Plasmids expressing ParC and ParE.Vent DNA polymerase was used to amplify the parC and parE genes from 7785 DNA in the presence of 1.5 mM MgCl2. For amplification ofparC, the forward primer (N6894) was 5′-TGGGCTTTGTATCATATGTCTAAC (artificialNdeI site underlined) and the reverse primer (VPC3) was 5′-CATTTCTCGAGTTTATCTTCAGTAACTAC [XhoI site underlined; converts TAA termination codon to CTC(Leu)]. For parE, the forward primer (N7043) was 5′-AGGAGGTTCCATATGTCAAAAAAGG (artificialNdeI site converts GTG initiation codon to ATG), and the reverse primer (N7044) was 5′-TATTTGGATCCATTAAACACTGTC (BamHI site underlined), which corresponds to sequence downstream of the natural stop codon. The PCR conditions were denaturation at 94°C for 1 min, annealing at 45°C (forparC) or 50°C (parE) for 1 min, and polymerization at 72°C for 3 min. Reactions were performed over 30 cycles. The resulting 2.5-kb parC and 1.9-kb parEproducts were digested with NdeI and XhoI andNdeI and BamHI, respectively. DNA fragments were purified from low-gelling temperature agarose, recovered, and ligated into appropriately cut kanamycin-resistance plasmid pET-29a (forparC) and ampicillin-resistance plasmid pET-19b (forparE), yielding pXP13 and pXP14, respectively.
Protein overexpression and purification.Plasmids pXP9, pXP10, pXP13, and pXP14 were transformed separately into E. coli BL21(λDE3)pLysS. GyrA and GyrB proteins were purified by the same procedure. Single colonies of BL21 containing pXP9 or pXP10 were picked from plates and grown overnight at 37°C in 50 ml of Luria-Bertani (LB) medium containing the selective antibiotic. A culture (15 ml) of the overnight growth was used to inoculate 750 ml of LB medium containing ampicillin (100 μg/ml) or kanamycin (50 μg/ml). Cells were grown at 37°C for about 3 h, until the optical density at 600 nm reached 0.4 to 0.6. IPTG was added to a final concentration of 1 mM, and growth was continued for a further 3 h. Bacteria were harvested by centrifugation at 5,000 × g for 12 min at 4°C. The supernatant was discarded, and the bacterial pellet was resuspended in 10 ml of 20 mM Tris-HCl (pH 7.9), 500 mM NaCl, and 5 mM imidazole prior to flash freezing in liquid nitrogen and storage at −70°C overnight. The suspension was thawed on ice, and lysozyme and Brij were added to final concentrations of 0.02 and 0.12%, respectively. Incubation was continued on ice for another 30 min, and then the mixture was centrifuged at 35,000 × g for 60 min. The supernatant was carefully removed to a sterile precooled 50-ml tube (Falcon) and mixed with 3 ml of 50% Ni-nitrilotriacetic acid (NTA) resin slurry (Qiagen). The tube was gently agitated at 4°C overnight. The mixture was poured into a column and washed initially with 10 ml of 20 mM Tris-HCl (pH 7.9), 500 mM NaCl, and 5 mM imidazole, followed by 10 ml of a solution containing 20 mM Tris-HCl (pH 7.9), 500 mM NaCl, and 60 mM imidazole. The column was then washed with 10 ml of a mixture of 20 mM Tris-HCl (pH 7.9), 1.5 M NaCl, and 60 mM imidazole. The histidine-tagged GyrA and GyrB proteins were eluted with 10 ml of 20 mM Tris-HCl (pH 7.9), 500 mM NaCl, and 250 mM imidazole. The protein fractions were examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and those containing GyrA or GyrB proteins were pooled (total volume of 2 to 3 ml) and dialyzed overnight at 4°C against 2 liters of 50 mM Tris-HCl (pH 7.9), 200 mM NaCl, and 30% glycerol. The protein solution was spun at 35,000 × g for 30 min at 4°C to remove precipitate. The supernatant was transferred to a fresh precooled tube, and dithiothreitol (DTT) and EDTA were added to final concentrations of 5 and 1 mM, respectively. The GyrA and GyrB proteins were then flash frozen in aliquots in liquid nitrogen and stored at −70°C. Approximately 2 mg of highly purified GyrA or 8 mg of purified GyrB was obtained from 5-liter cultures of induced cells.
Methods for protein induction and purification were the same for both ParC and ParE. Growth of BL21(λDE3) pLysS transformed with pXP13 or pXP14, IPTG induction of cultures, and harvesting of cells and their lysis with lysozyme and Brij were performed as described for GyrA and GyrB. The crude cell extracts were centrifuged at 35,000 × g for 60 min, the supernatant was removed, and to the pellet (containing ParC or ParE as inclusion bodies) was added 10 ml of buffer A (20 mM Tris-HCl [pH 7.9], 500 mM NaCl, 5 mM imidazole) containing 6 M urea. The pellet was resuspended and left on ice for 30 min before centrifugation at 35,000 × g for 60 min. The supernatant was carefully removed to a 50-ml sterile tube and mixed with 1.5 ml of 50% Ni-NTA resin slurry (Qiagen). The tube was shaken gently overnight at 4°C. The mixture was then loaded into a column and washed initially with 10 ml of buffer A containing 6 M urea. The ParC (or ParE) protein was then renatured on the column by gradually reducing the urea concentration to zero. This was achieved at 4°C over a 10-h period by running a 300-ml linear gradient starting at 6 M urea in the wash buffer (20 mM Tris-HCl [pH 7.9], 500 mM NaCl, 20 mM imidazole). The His-tagged ParC or ParE protein was eluted with buffer containing 20 mM Tris-HCl (pH 7.9), 500 mM NaCl, and 250 mM imidazole. Column fractions were examined by SDS-PAGE, and those containing purified protein were pooled and dialyzed overnight against 2 liters of 50 mM Tris-HCl (pH 7.5), 200 mM NaCl, and 30% glycerol. The dialyzed solution was spun at 35,000 × g for 30 min at 4°C and transferred to a precooled tube, and DTT and Na3EDTA were added to final concentrations of 5 and 1 mM, respectively. Purified ParC and ParE proteins were flash frozen as aliquots in liquid nitrogen and stored at −70°C. From 5-liter cultures of transformed BL21 cells, 6 mg of ParC and 5 mg of ParE were recovered.
Topoisomerase catalytic and DNA cleavage assays.DNA supercoiling activity, reconstituted with purified S. pneumoniae GyrA and GyrB proteins, was assayed as previously described (22, 23) with relaxed pBR322 DNA (0.4 μg) as a substrate (total volume, 35 μl). For decatenation assays, the standard reaction mixture (20 μl) contained 40 mM Tris-HCl (pH 7.5), 6 mM MgCl2, 10 mM NaCl, 10 mM DTT, 200 mM potassium glutamate, 1 mM ATP, 50 μg of bovine serum albumin per ml, 450 ng of kDNA, and various amounts of ParC and ParE proteins. Reaction mixtures were incubated at 37°C for 1 h, the reactions were terminated by addition of dye mix, and then the products were analyzed by electrophoresis in 1% agarose.
Relaxation assay mixtures (20 μl) containing 40 mM Tris-HCl (pH 7.5), 6 mM MgCl2, 10 mM NaCl, 1 mM spermidine, 1 mM ATP, 50 μg of bovine serum albumin per ml, 450 ng of supercoiled pBR322, and topoisomerase IV subunits were incubated at 37°C for 1 h. Reactions were terminated, and DNA products were examined by electrophoresis on 1.2% agarose gels.
DNA cleavage assays were carried out as for DNA supercoiling, except ATP was omitted, and relaxed-DNA substrate was substituted for with supercoiled pBR322 DNA. In each reaction, GyrA (0.45 μg) and GyrB (1.7 μg), or purified ParC (0.45 μg) and ParE (1.7 μg) proteins, were incubated with DNA in the absence or presence of fluoroquinolone drug for 1 h at 25°C. Three microliters of 2% SDS and 3 μl of a 1-mg/ml concentration of proteinase K were added, and incubation continued for 30 min at 37°C. Reactions were stopped by adding 7 μl of dye mix, and samples were analyzed by electrophoresis in a 1% agarose gel run at 3.5 V/cm. DNA was visualized by staining with ethidium bromide, and the gel was photographed under UV transillumination with a Land camera and Polaroid 665 film. The extent of DNA cleavage was quantitated from photographic negatives with a Molecular Dynamics personal densitometer SI and ImageQuant software.
Nucleotide sequence accession number.The S. pneumoniae 7785 genomic DNA sequence contained in plasmids pXP11 and pXP12 has been submitted to the EMBL Data Library under accession no. AJ005815.
RESULTS
Overexpression of S. pneumoniae GyrA and GyrB proteins and reconstitution of gyrase activity.Our initial attempts to purify the native S. pneumoniae GyrA and GyrB proteins by using inducible plasmid constructs in E. coli were frustrated by the low levels of expression, particularly for GyrA (32a). To circumvent this difficulty and to facilitate purification, we engineered recombinant genes that express GyrA and GyrB proteins carrying histidine tags at their C-terminal and N-terminal ends, respectively. The S. pneumoniae gyrA andgyrB genes were each amplified by PCR with S. pneumoniae 7785 genomic DNA as a template and Vent DNA polymerase, which has proofreading activity, thereby minimizing the introduction of PCR errors. The full-length gyrB gene contained betweenNdeI and XhoI sites was inserted in frame into expression vector pET-19b downstream of a T7 promoter, resulting in plasmid pXP9, which was transformed into E. coliBL21(λDE3)pLysS. Induction by IPTG of the T7 polymerase gene on the λDE3 prophage was expected to produce a recombinant 72-kDa GyrB protein carrying a decahistidine tag at its N-terminal end. The gyrA expression construct pXP10 was obtained by inserting the gyrA gene as two PCR fragments into plasmid pET-29a by creating an XhoI site 3′ of the gene that altered the Ala-Stop gyrA codons to Leu-Glu. Expression of the gene was expected to produce the full-length 97-kDa GyrA protein carrying a hexahistidine tag at its C-terminal end. Milligram amounts of soluble His-tagged GyrA and GyrB proteins that were each >90% homogeneous were isolated from cleared lysates of induced BL21 transformants by one-step nickel chelate chromatography (Fig.1A).
SDS-PAGE analysis of purified S. pneumoniaegyrase (A) and topoisomerase IV (B) subunits. Lanes: A and B, GyrA and GyrB proteins, respectively; C and E, ParC and ParE proteins, respectively. Approximately 2 μg of each protein sample was loaded on an SDS–7.5% polyacrylamide gel, and, following electrophoresis, proteins were revealed by staining with Coomassie blue. Sizes of protein markers (M) are indicated to the left.
The recombinant gyrase proteins were tested for DNA supercoiling activity with relaxed pBR322 DNA as a substrate (Fig.2A). Neither subunit alone induced DNA supercoiling in the absence or presence of 1.4 mM ATP (Fig. 2A, lane A or B). The combination of GyrA and GyrB subunits led to plasmid supercoiling in the presence of ATP that was not observed when ATP was omitted (Fig. 2A, lanes AB). The specific activities of the GyrA and GyrB proteins in the supercoiling assay were each 2 × 105 U/mg.
Enzymatic activities of S. pneumoniae gyrase and topoisomerase IV subunits. (A) S. pneumoniae GyrA and GyrB proteins reconstitute an ATP-dependent DNA supercoiling activity. Relaxed plasmid pBR322 substrate (0.4 μg) was incubated with gyrase proteins in the absence or presence of 1.4 mM ATP. Reactions were stopped, and the DNA was examined by electrophoresis in 1% agarose. Lanes: a, supercoiled pBR322 control; b, relaxed DNA control; A, relaxed DNA and GyrA protein (20 ng); B, relaxed DNA and GyrB (20 ng); AB, relaxed DNA and both GyrA (20 ng) and GyrB (20 ng). N, R, and S denote nicked, relaxed, and supercoiled DNA, respectively. (B) Decatenation of kDNA by S. pneumoniae topoisomerase IV. kDNA (0.4 μg) was incubated with purified ParC (2 ng) and/or ParE (20 ng) in the presence or absence of 1 mM ATP prior to agarose gel electrophoresis. Lanes a, kinetoplast DNA control; C, kDNA and ParC; E, kDNA and ParE; CE, kDNA with ParC and ParE. kDNA remains in the wells of the agarose gel; the position of monomer circles is indicated.
Several considerations indicate the observed gyrase activity arises from the recombinant proteins and is not due to copurifying enzyme from the E. coli host. First, we would not expect the E. coli gyrase subunits to be retained on the nickel column, because they lack histidine tags. Second, protein binding and elution during chromatography were carried out in the presence of 1.5 M NaCl, which should disrupt any heterologous subunit interactions. Studies ofE. coli gyrase bound via GyrB to novobiocin-Sepharose have shown that 1.0 M NaCl allows selective elution of GyrA (37). Third, SDS-PAGE analysis did not reveal the presence of bands that would correspond to the 100-kDa GyrA and 90-kDa GyrB subunits fromE. coli (Fig. 1A). Fourth, neither subunit alone was able to supercoil DNA in the presence of ATP (Fig. 2). Finally, although DNA supercoiling by E. coli gyrase is inhibited by ciprofloxacin with a 50% inhibitory concentration (IC50) of 0.7 μM (27a), the ciprofloxacin IC50 for the recombinant enzyme was 60-fold higher (described below), a value similar to that measured for gyrase from other gram-positive species (e.g., S. aureus) (4, 38).
Expression, purification, and characterization of S. pneumoniae topoisomerase IV.Full-length parC andparE genes were amplified by PCR with Vent DNA polymerase and the S. pneumoniae 7785 DNA template and inserted into pET vectors in a similar manner to that described for the gyrase genes, yielding expression constructs pXP13 and pXP14, respectively. Induction of exponentially growing cultures of BL21(λDE3) transformants resulted in low-level expression of ParC and ParE, both predominantly in an insoluble form (not shown). However, solubilization in 6 M urea and on-column renaturation yielded milligram amounts of soluble 93-kDa ParC and 72-kDa ParE proteins that were both >95% homogeneous by SDS-PAGE (Fig. 1B).
Enzyme activity was examined with a decatenation assay that monitors the ATP-dependent unlinking of DNA minicircles from kDNA (Fig. 2B). Neither the ParC nor ParE subunit alone, assayed in the absence or presence of 1 mM ATP, had decatenation activity: the kDNA remained intact and failed to migrate from the wells. However, when combined, ParC and ParE promoted decatenation and minicircle release in a reaction dependent on ATP. By using an excess of the complementing subunit, the specific activities of the S. pneumoniae ParC and ParE proteins in the decatenation assay were determined as 106 and 105 U/mg, respectively.
ATP-dependent relaxation of supercoiled plasmid pBR322 by topoisomerase IV was also examined (data not shown). Both subunits and ATP were needed to reconstitute DNA relaxation. The specific activities of the ParC and ParE proteins in the relaxation assay were 2.7 × 104 and 1 × 104 U/mg, respectively. Thus, the purified S. pneumoniae ParC and ParE proteins reconstituted highly efficient DNA decatenation and relaxation commensurate with a DNA topoisomerase IV activity.
Inhibition of gyrase and topoisomerase IV enzyme activities by antipneumococcal fluoroquinolones.Access to S. pneumoniae topoisomerases allowed us to examine and compare the effects of various fluoroquinolones on their principal enzymatic activities: DNA supercoiling for gyrase and decatenation by topoisomerase IV (Fig. 3 and 4). For gyrase assays, 1 U of enzyme was employed, which in the absence of drug is sufficient to convert 50% of 0.4 μg of relaxed plasmid pBR322 DNA to the supercoiled form under standard conditions (Fig. 3, lane 3). DNA supercoiling was inhibited in a dose-dependent manner by ciprofloxacin, sparfloxacin, and clinafloxacin, with IC50s (the concentration of drug required to inhibit DNA supercoiling by 50%) of 40, 40, and 2.5 μM, respectively. Thus, ciprofloxacin and sparfloxacin were equally good inhibitors, but were 16-fold less potent than clinafloxacin.
Fluoroquinolone inhibition of DNA supercoiling byS. pneumoniae gyrase. Relaxed pBR322 DNA (0.4 μg) was incubated with GyrA (1 U) and GyrB (1 U) proteins plus 1.4 mM ATP in the absence or presence of drugs. DNA was analyzed as described in the legend to Fig. 2. The concentration of ciprofloxacin (CIP), sparfloxacin (SPAR), or clinafloxacin (CLN) included in each reaction is shown on the figure. Lanes a and b, supercoiled- and relaxed-DNA controls, respectively.
In the case of DNA decatenation by S. pneumoniaetopoisomerase IV, reactions were set up with 1 U of topoisomerase IV activity in the absence or presence of increasing drug concentrations (Fig. 4). In the absence of drug, the enzyme converted approximately 50% of the input kDNA to free minicircles. Inclusion of any of the three quinolones resulted in a dose-dependent inhibition of decatenation. Ciprofloxacin and sparfloxacin were comparably effective, with IC50s (the drug concentration that inhibits decatenation by 50%) of 10 to 20 μM (Fig. 4). Clinafloxacin was the best inhibitor, displaying an IC50 of 1 to 2.5 μM.
Inhibition of DNA decatenation by S. pneumoniae topoisomerase IV. kDNA (0.4 μg) was incubated with ParC (1 U), ParE (1 U), and 1.4 mM ATP in the presence or absence of quinolones (ciprofloxacin [CIP], sparfloxacin [SPAR], or clinafloxacin [CLN]) at the indicated concentrations. DNA was analyzed by agarose gel electrophoresis.
Fluoroquinolone stabilization of cleavable complexes: preferential DNA breakage by S. pneumoniae topoisomerase IV.Figure5 compares the abilities of different fluoroquinolones to induce DNA linearization of supercoiled pBR322 DNA by S. pneumoniae gyrase and topoisomerase IV. Supercoiled pBR322 DNA was incubated with enzyme in the absence or presence of quinolone. DNA breakage was induced by addition of SDS, and following proteinase K digestion, the DNA was examined by agarose gel electrophoresis. Ciprofloxacin and sparfloxacin were comparably efficient at promoting gyrase-mediated DNA breakage, with the proportion of linear DNA increasing in a dose-dependent fashion (Fig.5A). For these quinolones, the (CC25) concentration of drug necessary to promote 25% linearization of the DNA by gyrase was 80 μM. However, clinafloxacin was much more effective in promoting DNA breakage, with a CC25 of 2.5 μM (Fig. 5A). At >5 μM, linear DNA was the predominant DNA species (Fig. 5A). Thus, clinafloxacin was some 20- to 40-fold more effective than either ciprofloxacin or sparfloxacin in mediating cleavable complex formation with gyrase.
Fluoroquinolone-mediated DNA cleavage is more efficient for S. pneumoniae topoisomerase IV than gyrase. (A) Drug-dependent DNA breakage by gyrase. Supercoiled pBR322 (0.4 μg) was incubated with pneumococcal GyrA (25 U) and GyrB (100 U) proteins at the concentrations of ciprofloxacin, (CIP), sparfloxacin (SPAR), and clinafloxacin (CLN) indicated on the figure. After treatment with SDS and proteinase K, DNA samples were examined by electrophoresis in 1% agarose. Lanes a and b, supercoiled and linear pBR322 DNA, respectively. (B) DNA cleavage by topoisomerase IV. Amounts of ParC and ParE equivalent to those of the gyrase subunits in panel A were incubated with various quinolones and pBR322 DNA. After induction of DNA breakage, DNA samples were processed for agarose gel electrophoresis as described for gyrase.
The results of DNA cleavage experiments involving topoisomerase IV are shown in Fig. 5B. The same molar amounts of topoisomerase IV proteins were employed as were used for gyrase in Fig. 5A. It can be seen immediately that much lower concentrations of each of the three quinolones were needed to induce DNA breakage by topoisomerase IV: the CC25s for ciprofloxacin, sparfloxacin, and clinafloxacin were 1.0, 1.0, and 0.1 μM, respectively. Table1 collects together the various IC50 and CC25 data for comparison with the quinolone MICs previously determined for S. pneumoniae 7785. These data are representative and have been obtained reproducibly in several independent experiments. Clinafloxacin was the most potent enzyme inhibitor against gyrase and topoisomerase IV.
Inhibitory effects of fluoroquinolones on S. pneumoniae gyrase, topoisomerase IV, and growth of S. pneumoniae 7785
DISCUSSION
We have expressed S. pneumoniae gyrase and topoisomerase IV in E. coli, purified the enzymes to homogeneity, and compared their interactions with new antipneumococcal quinolones. The key to expressing the S. pneumoniae gyrase and topoisomerase IV proteins was the use of His-tagged vectors. Previously, we had not been successful in purifying either enzyme in native form from S. pneumoniae or in isolating native recombinant S. pneumoniae proteins expressed in E. coli, due to their low level of expression, particularly for GyrA (32a). The use of His-tagged subunits facilitated rapid protein purification by metal affinity chromatography, yielding products free of host gyrase and topoisomerase IV proteins. Purification of native type II topoisomerases from S. pneumoniae has not been reported. However, the pneumococcal gyrase and topoisomerase IV subunits described here were comparably active to their counterparts from other sources. Thus, the S. pneumoniae GyrA and GyrB proteins had specific activities in the DNA supercoiling assay of ∼2 × 105 U/mg, which compare favorably with specific activities of 106 and 105 U/mg observed for native E. coli GyrA and GyrB (22) and 102 to 103 U/mg measured for the native and recombinant S. aureus gyrase proteins, respectively (4, 38). The activity of S. aureus gyrase could be increased 500-fold by the inclusion of 700 mM potassium glutamate, perhaps mimicking the high intracellular aspartate and glutamate levels found in S. aureus(4). (We found that addition of 700 mM potassium glutamate increased the activity of S. pneumoniae gyrase approximately 15-fold [32a]). Although we do not have wild-typeS. pneumoniae proteins for analysis, from these results, it appears that the histidine-tagged S. pneumoniae gyrase proteins display comparable activity to the native E. coliand S. aureus counterparts. Moreover, it has been reported recently that S. aureus gyrase and topoisomerase IV holoenzymes reconstituted from His-tagged subunits have specific activities identical to published values, and these proteins behaved similarly to the native enzymes in their responses to quinolones (35).
For S. pneumoniae, there are a number of interesting features of the in vitro and in vivo responses to quinolones that deserve comment (Table 1). First, sparfloxacin and ciprofloxacin had very similar inhibitory activities in vitro, whether measured in the DNA supercoiling, DNA decatenation, or DNA breakage assays. Despite these similarities, comparison of MICs shows that sparfloxacin was fourfold more potent than ciprofloxacin against S. pneumoniae 7785 (Table 1). One factor that could plausibly explain the greater in vivo activity of sparfloxacin is less-efficient efflux compared to that of ciprofloxacin. Second, a striking feature of Table1 is that clinafloxacin, when examined in any of the enzyme assays, was a markedly more potent inhibitor than sparfloxacin or ciprofloxacin. Thus, the IC50s of clinafloxacin for DNA supercoiling by gyrase and for DNA decatenation by topoisomerase IV were some 10- to 20-fold lower. Moreover, the CC25s obtained in the DNA cleavage experiments were 32-fold lower for gyrase and 10-fold lower for topoisomerase IV. The CC25s are useful, because trapping of type II topoisomerases by quinolones as cleavable complexes on DNA is thought to be the cytotoxic lesion initiating the antibacterial action of these drugs (18, 19). The much lower values for clinafloxacin support our recent suggestion that the greater antibacterial activity of the drug is due to intrinsic tight binding to its enzyme targets in vivo (32). However, the 10- to 32-fold differences in CC25s over those of sparfloxacin and ciprofloxacin are not reflected proportionally in the MICs (Table 1), suggesting that other factors contribute to drug action in vivo. Third, it is interesting that all three quinolones showed an in vitro preference for topoisomerase IV, with CC25s that were 25- to 80-fold lower than those against gyrase (Table 1), the inverse of that seen with E. coli gyrase and topoisomerase IV. This was particularly surprising, because genetic experiments have shown that the three quinolones have different targets in S. pneumoniae: ciprofloxacin and sparfloxacin act preferentially through topoisomerase IV and gyrase, respectively, whereas clinafloxacin acts through both, with a weak preference for gyrase (32). (A recent abstract has also reported the preferential targeting of topoisomerase IV by sparfloxacin in vitro and the equipotency of sitafloxacin [24].) Overall, there was no rigid correlation between enzyme inhibition, MICs, and target preferences in vivo.
Recently, the preference of quinolones for topoisomerase IV has also been observed in comparisons of S. aureus topoisomerase IV and gyrase in vitro, using either native or His-tagged holoenzymes (4, 35). Thus, ciprofloxacin, sparfloxacin, and norfloxacin were some 10- to 250-fold more effective in stabilizing covalent complexes with native S. aureus topoisomerase IV than with DNA gyrase (4). Similarly clinafloxacin, trovafloxacin, ciprofloxacin, norfloxacin, and oxolinic acid showed a 2- to 50-fold preference for His-tagged topoisomerase IV (35). Although for a number of quinolones, these results concur with the known in vivo target preference for topoisomerase IV, in the case of sparfloxacin, there are differences between the in vivo and in vitro data. Sparfloxacin is >50-fold more effective in stimulating DNA cleavage by topoisomerase IV over gyrase in vitro (4), and yet it appears to kill S. aureus cells by acting through both gyrase and topoisomerase IV (8). Clearly, for both S. aureus and S. pneumoniae, there are differences between targeting preferences seen in vitro with purified proteins and those suggested by genetic approaches.
How can these differences be reconciled? One hypothesis would be to assume that all quinolones preferentially target DNA topoisomerase IV in S. pneumoniae. For those quinolones such as sparfloxacin and clinafloxacin that select gyrA (but not parCor parE) QRDR mutants in the first step, it would then have to be assumed that these are not genuine single-step mutants, but carry a second mutation (thus far undetected) in topoisomerase IV lying outside of the QRDRs. However, to explain the lack of cross-resistance with other quinolones, this putative second mutation would have to affect specifically sparfloxacin and clinafloxacin and not, e.g., ciprofloxacin. Second, the observed frequencies of first-stepgyrA mutants obtained by challenge with sparfloxacin (at 4 × MIC) and clinafloxacin (at the MIC) are each in the range 5 × 10−10 to 8 × 10−10, values that appear very high for single-step selection of putative double mutants (31, 32). Third, from studies with E. coli and other gram-negative bacteria, it is known that gyrase is the in vivo and in vitro target of a variety of quinolones, including ciprofloxacin. Thus, either gyrase or topoisomerase IV can be a primary quinolone target. Although full sequence analysis of topoisomerase IV genes in quinolone-resistant mutants would be desirable, these constraints would seem to argue against the hypothesis favoring topoisomerase IV as the invariant quinolone target in S. pneumoniae.
The converse explanation is that measurements of cleavable complex formation using purified enzymes in vitro may not faithfully reflect the situation inside bacteria. First, it is possible that recombinant enzymes prepared by expression in E. coli may not reproduce the characteristics of the native proteins. For example, in our case, the renaturation of topoisomerase IV subunits from inclusion bodies, which was not required for gyrase subunits, may have affected the sensitivity of the reconstituted topoisomerase IV to quinolones. Although unlikely, such explanations remain to be formally excluded. Second, the in vivo targeting of gyrase by clinafloxacin and sparfloxacin in S. pneumoniae could result from preferential cleavable complex formation with gyrase (instead of topoisomerase IV) under intracellular conditions. Obviously, the particular DNA template, salt, Mg2+, polyamine, ATP, DNA supercoiling, enzyme, and other conditions that prevail in the bacterium may be difficult to reproduce in vitro. It is also conceivable that individual quinolones are uniquely and unevenly distributed in the bacterial cell in a way that could affect complex formation with the two enzymes. This could be important if the target enzymes are themselves distributed in a nonuniform fashion. In fact, it is already known from studies withBacillus subtilis that topoisomerase IV has a bipolar localization, whereas gyrase is associated with the nucleoid (14). More data will be needed about cleavable complex formation in S. pneumoniae.
Finally, and our preferred model, the discrepancy between in vitro and in vivo results could arise through differential lethality of cleavable complexes in vivo. Obviously, unlike the assay of DNA breakage in vitro, which largely reflects drug-target affinity, in vivo targeting revealed through genetic experiments identifies the most important killing pathway, in which killing is a complex process involving drug-enzyme binding and formation of a cleavable complex, followed by downstream events, such as collisions with replication forks that convert the cleavable complex into a lethal lesion thought to be a chromosomal double-stranded DNA break. It has been suggested from studies of E. coli that gyrase acts ahead of the replication fork, whereas topoisomerase IV acts predominantly behind the fork, allowing time for repair of quinolone-induced DNA damage (44). This scenario predicts that cleavable complex formation through gyrase would give rapid killing, whereas that involving topoisomerase IV would be a slow process, a pattern of behavior that has recently been observed in comparisons of norfloxacin’s actions against E. coli and itsgyrA mutants (18). However, unlike E. coli, in which gyrase is usually the primary quinolone target, the situation in S. pneumoniae is more complex, in that different quinolones appear to have different targets in vivo. Conceivably, cleavable complex formation through topoisomerase IV or gyrase may be more or less lethal for some quinolones than for others. Resolution of these questions must await the development of an in vitroS. pneumoniae replication model in which quinolone effects can be examined in detail. However, the efficient expression system forS. pneumoniae type II topoisomerases described in this paper should facilitate further studies of the selectivity of antipneumococcal fluoroquinolones and open the way to crystallographic approaches aimed at elucidating the molecular basis of quinolone-topoisomerase interactions.
ACKNOWLEDGMENTS
We thank Howard Nash for useful discussions; Ming-shi Li for help with laser densitometry; and Stephen J. Gracheck, Michael A. Cohen, and Jing Li for helpful comments on the manuscript.
This work was supported by a grant from Parke-Davis, Co.
FOOTNOTES
- Received 4 January 1999.
- Returned for modification 4 February 1999.
- Accepted 3 March 1999.
- Copyright © 1999 American Society for Microbiology
