|Home | About | Journals | Submit | Contact Us | Français|
Antibacterial compounds with new mechanisms of action are needed for effective therapy against drug-resistant pathogens in the clinic and in biodefense. Screens for inhibitors of the essential replicative helicases of Bacillus anthracis and Staphylococcus aureus yielded 18 confirmed hits (IC50 ≤ 25 μM). Several (5 of 18) of the inhibitors were also shown to inhibit DNA replication in permeabilized polA-deficient B. anthracis cells. One of the most potent inhibitors also displayed antibacterial activity (MIC ~5 μg/ml against a range of Gram-positive species including bacilli and staphylococci) together with good selectivity for bacterial vs. mammalian cells (CC50/MIC >16) suitable for further optimization. This compound shares the bicyclic ring of the clinically proven aminocoumarin scaffold, but is not a gyrase inhibitor. It exhibits a mixed mode of helicase inhibition including a component of competitive inhibition with the DNA substrate (Ki = 8 μM) and is rapidly bactericidal at 4× MIC.
The increasing prevalence of antibiotic-resistant bacterial strains continues to erode the efficacy of current antibiotics. The development of new antibacterials that are not subject to existing mechanisms of resistance offers a solution to this growing medical need, but successful discovery and development of new antibacterials have been difficult to achieve.1 The bacterial DNA replication complex presents an attractive, and largely unexploited, target for development of new antibacterials. Among inhibitors which target bacterial DNA replication, only the fluoroquinolones, inhibitors of bacterial gyrase and topoisomerase, are in current clinical use.2 Potent inhibitors of Gram-positive DNA polymerase have also been described but have not been tested in the clinic yet.3 The bacterial replicative DNA helicase is a promising target for new antibiotic discovery because it fulfills an essential role in DNA replication and exhibits no significant homology to mammalian helicases. Through the ATP-dependent unwinding of the DNA duplex, replicative helicases allow access by the rest of the replication machinery to the replication fork and thus permit duplication of the bacterial genome. These enzymes function as hexameric rings, with the DNA occupying the central channel of the hexamer.4 Bacterial replicative helicases have been demonstrated to be essential for bacterial growth, and inhibitors would likely be bactericidal since existing gyrase and polymerase inhibitors are known to kill susceptible bacteria. Bacterial cells contain a variety of putative or actual helicases in addition to the replicative enzyme, but they are not closely related structurally to the replicative helicases and most are involved in DNA repair or plasmid replication and are not essential for bacterial growth.
Prior screening for inhibitors of bacterial replicative helicases has yielded very few potent, selective and non-cytotoxic inhibitors. In fact, only two compounds have been described as bacterial replicative helicase inhibitors -- the natural product flavonoid myricetin inhibits Escherichia coli DnaB helicase5, and a triaminotriazine was identified in a screen for inhibitors of Pseudomonas aeruginosa DnaB helicase.6 However, these compounds also exhibit significant cytotoxicity in mammalian cell culture. Because orthologous antibacterial targets from different species contain subtle sequence differences which could make them more accessible to small molecule inhibitors,7, 8 we undertook the screening of compounds for inhibition of the replicative helicases from two additional species, Staphylococcus aureus and Bacillus anthracis. Surprisingly, the hit rates for the two helicases were remarkably different, and with few exceptions, hits were more potent vs. the B. anthracis helicase than they were vs. the S. aureus helicase, regardless of the screen in which they were first identified as helicase inhibitors. One of the most potent and selective inhibitors discovered (see compound 2 below) shares a portion of the aminocoumarin chemotype. While bioactive compounds with a coumarin scaffold have been known for decades, they are inhibitors of DNA gyrase and are structurally distinct from the coumarins described here.9 Nevertheless, favorable clinical history with this class of compounds suggests that further development of coumarin-type helicase inhibitors is feasible.
Genes for the B. anthracis and S. aureus replicative helicases were cloned and expressed in E. coli, the enzymes were purified, and fluorescence-based assays were developed for each and optimized as high throughput screens. A total of over 186,000 small synthetic molecules were screened to identify inhibitors of the S. aureus or B. anthracis helicase-catalyzed strand unwinding reaction. Primary hits were selected and confirmed by re-assay, requiring over 50% inhibition in at least two of three replicates. The overall confirmed hit rate was about 0.08%, but when calculated separately for each helicase, it was nearly 10-fold higher for the B. anthracis enzyme than for the S. aureus enzyme (Table 1).
Confirmed inhibitors of each of the helicases were characterized further to eliminate false positives which act by mechanisms other than direct inhibition of helicase and to measure the concentration-dependence of helicase inhibition. First, hits were examined in an ethidium bromide displacement assay10 to eliminate compounds which inhibit strand unwinding by binding to the DNA duplex substrate rather than to the helicase. Second, hits were tested in a radiometric assay of helicase activity to ensure that hits block strand unwinding rather than simply quenching FAM fluorescence in the FRET assay. Several hits which resemble known intercalators or minor groove binders or were strong quenchers were eliminated by these secondary assays. Third, hits were tested for inhibition of Pseudomonas aeruginosa AmpC β-lactamase in the presence of various concentrations of Triton X-100 to detect compounds acting promiscuously by a colloidal aggregate mechanism.11 None of the confirmed hits exhibited inhibition of AmpC at 0.01% Triton X-100, the concentration used in the FRET helicase assays, indicating that aggregates are not likely to be responsible for the observed helicase inhibition. Finally, hits from each helicase screen were examined for inhibition of the helicase of the other species, and the concentration-dependence of inhibition (IC50) was determined.
About 10% of the 160 confirmed primary hits, a total of 18 compounds were validated by these secondary assays and exhibited concentration-dependent inhibition with IC50 values <25 μM vs. at least one of the two helicases (Table 1). These fall into five chemotypes with three additional compounds as singletons (Table 2). With the exception of one chemotype (see below), confirmed hits were demonstrated by LC-MS analysis to be of correct mass and sufficient purity (>95%) for further evaluation.
In order to evaluate the selectivity of the inhibitory effects of these compounds, all confirmed hits were tested for (a) potency of inhibition of DNA replication in permeabilized polA-deficient B. anthracis cells, (b) minimal inhibitory concentration (MIC) vs. growth of B. anthracis and S. aureus cells, and counter-screened for (c) potency of inhibition of the replicative DNA polymerase Pol IIIC of B. subtilis and (d) for effects on the viability of mammalian cells (cytotoxicity). Structures of the confirmed hits along with results of these confirmatory assays are shown in Table 2. All data are expressed in μM to facilitate comparisons.
Three hits from the B. anthracis helicase screen and two hits from the S. aureus helicase screen share a coumarin chemotype (Series A, compounds 1-5, respectively, of Table 2). These coumarin-type compounds are potent screening hits, with IC50 values ranging from 5-10 μM vs. B. anthracis helicase. All five compounds exhibit significant inhibition of both B. anthracis and S. aureus helicases, but they are all less potent vs. S. aureus helicase, with IC50's ranging from about 2 to 20-fold higher than those for B. anthracis helicase. They are poor inhibitors of another DNA replication enzyme, B. subtilis replicative DNA polymerase Pol IIIC (IC50's >100 μM) confirming their selectivity for helicase. As expected for helicase inhibitors, they inhibit DNA replication in permeabilized polA-deficient B. anthracis Sterne cells, but the potency of compounds 1, 4, and 5 is less than that exhibited in the in vitro FRET-based assay. Coumarins 2 and 4 are not cytotoxic to mammalian cells when tested up to 100 μM. Furthermore, compound 2 inhibits the growth of B. anthracis cells with an MIC of 6 μM, resulting in an overall selectivity index (CC50/MIC) >16.
Two hits from the B. anthracis helicase screen share a benzothiazole chemotype (Series B, compounds 6-7 of Table 2). They both exhibit IC50 values vs. B. anthracis helicase of 12 μM but are slightly less potent vs. S. aureus helicase. These two compounds exhibit no detectable inhibition of B. subtilis DNA polymerase Pol IIIC, but also fail to demonstrate detectable inhibition of DNA replication in permeabilized B. anthracis cells. Neither of the benzothiazoles has a detectable MIC, but some cytotoxicity is observed, with CC50's of 10 and 70 μM.
Four confirmed hits from the B. anthracis helicase screen share a rhodanine chemotype (Series C, compounds 8-10, Table 2). This class of hits is unique for two reasons -- they are the most potent helicase inhibitors identified, and all are approximately equipotent against both B. anthracis and S. aureus helicase, with IC50 values of 1-4 μM. Three members of this class exhibit no detectable cytotoxicity (CC50 >100 μM), but the replacement of the 4-chloro substituent with a nitro-group (compound 10) resulted in pronounced cytotoxicity (CC50 = 11 μM). The three non-cytotoxic rhodanines (8, 9 & 11) inhibit DNA replication in permeabilized B. anthracis cells, but with potencies considerably poorer than those demonstrated in direct helicase assays. None exhibited a detectable MIC.
Two confirmed hits from the B. anthracis helicase screen share a triazine chemotype (Series D, compounds 12 and 13, Table 2), and they bear some resemblance to the triaminotriazine inhibitor of P. aeruginosa replicative helicase described previously.6 They are moderately potent inhibitors of B. anthracis helicase, with IC50 values of 12 and 24 μM, and are about half as potent vs. S. aureus helicase. Compound 13 fails to inhibit DNA replication in permeabilized B. anthracis cells and is quite cytotoxic to mammalian cells. Compound 12 inhibits DNA replication in permeabilized B. anthracis cells and displays about equal potency for bacterial and mammalian cell growth inhibition.
One confirmed hit from the B. anthracis helicase screen and one confirmed hit from the S. aureus helicase screen (Series E, compounds 14 and 15, respectively, in Table 2) share an N-phenyl pyrrole chemotype. While they appeared to be quite potent, they were the only confirmed hits that failed to exhibit the expected mass in LC-MS analysis. Further studies demonstrated that fresh DMSO solutions of re-ordered compounds exhibited no detectable helicase inhibitory activity, but solutions aged for several days in DMSO, did exhibit potent activity. This class of inhibitors was considered artifactual due to instability and not pursued further. They have been described previously as inhibitors of other targets12, 13, and thus, appear to be promiscuous inhibitors as well (see Discussion).
Three confirmed B. anthracis helicase inhibitors (compounds 16, 17 and 18, Table 2) represent singleton chemotypes. Only compound 17 inhibited helicase in vitro as well as DNA replication in permeabilized cells. While not active against S. aureus helicase, this inhibitor exhibited an MIC against B. anthracis Sterne cells consistent with its IC50 and displayed low cytotoxicity, yielding a selectivity index (CC50/MIC) of ~2.7. Singleton 18 was about equally potent at inhibiting both B. anthracis and S. aureus helicases, but failed to produce a detectable MIC vs. either species. Computational searches of the screening library database revealed no analogs with 90% or better similarity to any of the three singletons (ChemBioOffice v 11.0, CambridgeSoft, Inc., MA).
Based on results from these secondary assays, three validated helicase inhibitors, compounds 2, 12, and 17, exhibit sufficient potency and selectivity to warrant further optimization. However, the coumarin-type inhibitor, compound 2, is clearly the most potent and selective (CC50>100 μM; helicase IC50 ~8 μM; MIC ~6-12 μM, or ~3-6 μg/ml, Table 2). Further characterization of this inhibitor, including mode of inhibition, determination of an inhibition constant (Ki), cidality, antibacterial spectrum, and preliminary SAR studies are described below.
Since the aminocoumarins coumermycin and novobiocin are known to inhibit the DNA gyrase function in DNA replication by competing with ATP binding 14, we examined the mode of inhibition of the coumarin-type helicase inhibitor compound 2. Assays of the ssDNA-stimulated ATPase activity of B. anthracis replicative helicase in the presence of compound 2 revealed only weak inhibition (16% inhibition in the presence of 100 μM compound 2), indicating that this coumarin-type helicase inhibitor has little or no effect on ATP binding or hydrolysis by helicase in the absence of its duplex DNA substrate. Helicase kinetic studies with varying concentrations of ATP substrate and inhibitor suggested that compound 2 is uncompetitive with ATP at low inhibitor concentrations (Fig. 1A). However, a Dixon plot analysis (1/V vs [I]) of the data was non-linear at inhibitor concentrations above the Ki at the two lower ATP concentrations tested, suggesting a possible second mode of inhibition at high compound 2 concentrations. Helicase kinetic studies with varying oligonucleotide substrate and inhibitor concentrations demonstrated that compound 2 exhibits a mixed competitive-uncompetitive mode of inhibition with the DNA substrate (Fig. 1B), and a Dixon plot indicates a Ki inhibition constant of ~8 μM (Fig. 1C). In addition, two coumarins (compounds 1 and 2) as well as one triazine (compound 12) helicase inhibitor were tested and failed to inhibit gyrase at concentrations of up to 50 μM (data not shown). This is not surprising since they lack the noviosyl sugar linkage, which is important for gyrase-inhibiting coumarins 9, 15. By contrast, two known gyrase-inhibiting coumarins, coumermycin and novobiocin, were potent inhibitors of gyrase in the assay (IC50's <1 μM) but did not inhibit B. anthracis helicase (data not shown). Thus, while these helicase inhibitors share a portion of the coumarin scaffold, they do not share a common biological mechanism.
Inhibitors of DNA replication such as the fluoroquinolones and the anilinouracils are typically bactericidal16, 17 because successful replication and segregation of the chromosome into each cell is essential for viability. The coumarin-type helicase inhibitor compound 2 was examined and proved to be rapidly bactericidal within 5.5 hr at 4× MIC (Fig. 2).
To probe the structure-activity relationship for compounds 1 and 2, nineteen additional structurally related compounds were examined against B. anthracis helicase (Table 3). The core bicyclic ring which is common to all five coumarin-type helicase inhibitors identified in screening was not altered, but some changes to all of the substituents were examined. The SAR results are summarized schematically in Fig 3A. Only one alteration increased the helicase inhibitory activity of 2, replacement of the substituted phenyl ring for an unsubstituted naphthyl ring. However, this change was also associated with increased cytotoxicity (compound 1, Table 2). Loss of the 8 position methyl group or shortening of the 3 position chain by one methylene group modestly reduced activity, but esterification of the 3 position carboxylic acid eliminated activity. These results suggest a pharmacophore representation as shown in Fig. 3B in which the acidic group and the oxygens interact with a hydrophilic region while the 4, 7, and 8 position substituents interact with a hydrophobic region.
The antibacterial spectra of the two related coumarin inhibitors, compounds 1 and 2, as well as that of the triazine, compound 12, were examined. The results reveal considerable breadth across the bacillus genus, with some limited activity against S. aureus (Table 4). Only the triazine displayed any activity against E. coli, and that was only in the tolC- efflux-deficient strain. The Gram-negative outer membrane barrier or efflux may rescue E. coli cells from inhibition by these compounds. These results suggest that helicase inhibitors could be chemically optimized to be effective on a range of Gram-positive species.
The results of this study have identified several new small molecule inhibitors of the replicative helicases of B. anthracis and S. aureus. Extensive characterization in secondary assays has also confirmed that they act by interacting with helicase rather than with the DNA substrate and that several display antibacterial activity with minimal cytotoxicity. These compounds may provide a starting point for further optimization and the development of antibacterials with utility in biodefense or clinical settings. Four aspects of this study are of special interest. First, very few compounds inhibit S. aureus replicative helicase as compared to the B. anthracis helicase. Second, the validated hits identified in this study have not been described previously as helicase inhibitors. Third, several inhibitors share a portion of the aminocoumarin scaffold but inhibit helicase specifically, with no activity against gyrase. Fourth, compounds with an N-phenyl pyrrole chemotype are surprisingly unstable and may be a source of artifactual inhibitory activity in many screens. These points are discussed below.
The relative difficulties in obtaining inhibitors of the S. aureus helicase and the differences in potency of inhibitors for the S. aureus vs. the B. anthracis helicases suggest that the B. anthracis helicase is more susceptible to small molecule inhibitors. The hit rate for B. anthracis helicase screening was nearly 10-fold higher than that for S. aureus helicase screening. While each screen was performed on a different library of compounds, the two libraries overlap substantially, and in addition, all hits are equipotent vs. the two helicases or more potent vs. B. anthracis helicase regardless of the screen from which they were identified. Structural differences in enzymatic targets from different bacterial species are known to affect the affinity for inhibitors significantly. In two recent examples, the affinities (measured as IC50 values) of enoyl-ACP reductase (FabI) and peptide deformylase (Def) for inhibitors varied up to 95-fold when measured for the E. coli ortholog as compared to the S. aureus ortholog.7, 8 At the level of primary amino acid sequence, the overall similarity between the B. anthracis and S. aureus replicative helicases is 76%, but the similarity between the two proteins is lower in the N-terminal domain (residues 1-152) and drops to 66%. While the ATP and DNA binding sites appear to be in the C-terminal domains, recent crystallographic evidence suggests that a single-stranded DNA binding site may reside in the N-terminal domain, and it may help guide the unwound strand to primase, which binds to the helicase N-terminal domain.4 Interestingly, one of the three arginines (residues 116, 117, and 120) comprising this site is substituted in the S. aureus primary sequence with glutamine (residue 120). Possibly this difference or some other difference between the sequences is responsible for the different affinities of the two helicases for small molecule inhibitors.
The DnaB-type replicative helicases from E. coli, S. aureus, and P. aeruginosa have been targeted previously in anti-infective screens. Screening assay read-outs have included electrochemiluminescence18, fluorescence or FRET6, 19, time-resolved FRET20, scintillation proximity (SPA)21, and radiometric detection of ATPase inhibition22, but few hits have been described and none have progressed further in drug development. A triaminotriazine structure was shown to inhibit P. aeruginosa DnaB, but it displays significant cytotoxicity.6 A large anti-bacterial screening effort undertaken by GlaxoSmithKline, Inc. produced no hits to S. aureus replicative helicase.1 While hits were obtained for another essential helicase in S. aureus, PcrA, which is involved in DNA repair and plasmid replication, no lead compounds could be developed from the hits.1 Inhibitors of the E. coli ortholog of PcrA, helicase IV, have also been described, but no information on cytotoxicity was provided and these inhibitors do not appear to have progressed further.22 Two investigators have described inhibition of E. coli helicases (DnaB and RepA) by the flavone myrecitin;5, 23 however, this compound is quite promiscuous and cytotoxic. Similarly, intercalators and minor groove binders which interact with DNA are potent helicase inhibitors, but they lack bacterial selectivity as well.24 Screening for inhibitors of B. anthracis replicative helicase as described in this report yielded new inhibitors not reported previously. Possibly this is a result of screening with a new ortholog of helicase having a subtly different binding site for inhibitors.
The most potent and selective inhibitor identified in this study, compound 2, contains a coumarin scaffold. A total of five coumarin-type helicase inhibitors were identified in the screen. They share a portion of a known bioactive chemotype, members of which inhibit gyrase by competing for ATP binding to the GyrB subunit. However, the helicase-inhibiting coumarins lack the noviosyl sugar moiety found in aminocoumarin antibacterials and consequently do not inhibit gyrase. This chemotype has a long history in antibacterial drug discovery and development. The aminocoumarin antibiotics novobiocin, clorobiocin and coumermycin A1 are produced by different Streptomyces strains and are potent antibacterials.25 Novobiocin was marketed for several years as an antibacterial in an oral formulation (“albamycin” by Pharmacia & UpJohn), but is no longer used clinically. Its loss of favor as a clinical therapeutic probably results from its weak activity against Gram-negatives and the development of target-based resistance.25-27 There were no serious toxicity issues with the use of novobiocin for acute infections despite the fact that it competes for an ATP binding site on gyrase. In fact, it is tolerated in humans at doses as high as 3g/day.28 This clinical history of the coumarin chemotype suggests that the coumarin helicase inhibitors could be well-tolerated in human therapy.
The only helicase inhibitors which failed our quality control analysis by liquid chromatography and mass spectrometry were the two members of the N-phenyl pyrrole chemotype. These two compounds were not active against helicase when freshly dissolved in DMSO, but gained activity with time in DMSO solution. It is important to note that the instability of these compounds may not be generally recognized and could be responsible for their reported inhibition of HIV type 1 entry (compound NB-213) and anthrax lethal factor (compound 1012). In light of our results, it is important to re-examine other putative activities of this class of compounds.
This first reported screen for inhibitors of B. anthracis replicative helicase has yielded several validated hits which have never been reported to inhibit helicase. The strong bias towards B. anthracis helicase in numbers of screening hits and potency of validated hits argues that this enzyme is more susceptible to inhibition by small molecules than is the S. aureus helicase. The most potent and selective inhibitor, compound 2, shares a portion of the aminocoumarin scaffold but inhibits helicase specifically.
The bacterial strains and plasmids used in this study, their sources and relevant genotypes are described in Table 5.
BHQ1 (“Hel-3′BHQ”) and 6-FAM (“Hel-5′FAM”) labeled 60-mer oligonucleotides (Table 6) were purchased from Operon, Inc. (Huntsville, AL) and Integrated DNA Technologies, Inc. (Coralville, IA), respectively, as HPLC purified. Slight variations in the mass quantitation and in RFU values required a careful calibration of each annealed batch to minimize batch-to-batch variation in the screen. The 30-mer capture strand (“Hel-Cap30”), which is complementary to the 5′-30 nucleotides of the FAM-labeled oligo, was purchased from Operon, Inc. as desalted, unpurified. The two labeled oligonucleotide strands were annealed at a 1:2 (Hel-5′FAM:Hel-3′BHQ) ratio prior to use in the FRET quenching helicase assay. PCR primers used in this study are also shown in Table 6.
The cloning and expression of the B. anthracis dnaB gene and subsequent purification of the helicase has been described.29 Briefly, the gene was amplified by PCR with primers HCASE45-5′and HCASE45-3′ (Table 6) and B. anthracis genomic DNA. The amplified gene was inserted into a pET30 vector (Novagen Inc., Madison, WI) under the control of a T7 promoter and confirmed by DNA sequencing. B. anthracis dnaB was over-expressed in E. coli strain BL21 (DE3) RIL (Stratagene Inc. CA) harboring the pET30-DnaBBA plasmid. Extraction of IPTG-induced cells was done as previously described (13). Protein was precipitated from the cell extract by addition of 0.25 g/ml (NH4)2SO4, resuspended in buffer A (25 mM Tris-HCl, (pH 7.5), 5 mM MgCl2, 10% glycerol, 5 mM DTT) and re-precipitated in 0.2 g/ml (NH4)2SO4. The protein pellet was resuspended in buffer A and fractionated by Q-Sepharose chromatography (GE Health sciences, Piscataway, NJ), which removed any contaminating endogenous E. coli helicase.30, 31 The flow through fractions were pooled and loaded onto a 6 ml S-Sepharose column equilibrated with buffer A100. B. anthracis helicase was eluted with a gradient of buffers A100 and A500. The peak fractions were identified by ssDNA dependent ATPase and DNA helicase activities in conjunction with SDS-PAGE. The active fractions were pooled and concentrated by ultrafiltration using a YM-30 membrane. The purified helicase was greater than 98% pure as analyzed by SDS-PAGE.29, 32 The enzyme was verified to be stable for several hours at room temperature, and for at least 12 weeks at−20°C in 20% glycerol storage buffer. B. anthracis helicase activity is dependent on the presence of ATP, with an optimum at 2.5 mM. The pH optimum is broad and centered around pH 7.8. Optimal concentrations of other assay components are as follows: 2 mM MgCl2, 0.5 mM DTT, 0.01% Triton X-100, and 25 mM NaCl.
The genes for the S. aureus replicative DNA helicase (dnaC) and helicase loader (dnaI) were amplified by PCR from genomic DNA isolated from S. aureus Smith using the primers 5′Sa-dnaC and 3′Sa-dnaC for the helicase gene and primers 5′Sa-dnaI and 3′Sa-dnaI for the loader gene (Table 6). Products were sequence-confirmed and cloned in the dual expression vector pET-Duet1 (Novagen), under control of a T7 promoter/lac operator. The helicase was expressed in native form, while the loader was expressed with an N-terminal hexahistidine tag to facilitate purification. Although the helicase loader proved unnecessary either for solubility or helicase activity, the dual expression clone was used since it produced larger quantities of helicase than did a clone containing the dnaC gene alone in the same vector. Following overnight induction by IPTG at 14°C, cells were harvested by centrifugation and lysed in a French press. The lysate was centrifuged, and the resulting cleared supernatant was precipitated by addition of 0.2 grams of (NH4)2SO4 per ml lysate. The precipate was redissolved in buffer B (50 mM Tris pH 7.5; 1M NaCl; 10% glycerol; 2 mM 2-mercaptoethanol) containing 1 mM PMSF, and applied to an IMAC-Ni++ column equilibrated in the same buffer. Following a wash in the same buffer, the column was eluted in a 0-200 mM imidazole gradient. The fractions containing helicase were pooled and applied to a phenyl sepharose column equilibrated in buffer B containing 1 mM EDTA; the column was eluted in gradient of decreasing NaCl (1 to 0 M) and increasing Triton X-100 (0 to 1%). The resulting helicase preparation was about 98% pure by SDS-PAGE and essentially free of nuclease activity as judged by minimal ATP-independent activity in the FRET assay. Optimal conditions for the reaction include a pH range of 7.6-8.4, a magnesium concentration of 2 mM, and an ATP concentration of 3 mM. The enzyme was stable at room temperature for at least two hours and at −20°C in 20% glycerol storage buffer for several months.
The FRET-based helicase activity assay was performed essentially as previously described6, 19 using labeled annealed oligodeoxynucleotides Hel-5′FAM:Hel-3′BHQ (Table 6). The assay is based on the helicase-mediated dissociation of two annealed oligonucleotides, one with a fluorescent label, the other bearing a quencher moiety. Two complementary annealed 60-mer oligonucleotides with noncomplementary 30-mer ends were labeled, respectively, with 6-carboxyfluorescein (6-FAM) and Black Hole Quencher 1 (BHQ1). When the two strands are unwound by helicase action, the 6-FAM emits an unquenched signal at 535 nm followed by excitation at 485 nm. To prevent the 6-FAM-labeled strand from reannealing with the quencher strand, an unlabeled 30-mer capture strand, Hel-Cap30, was added in 15-fold molar excess (Table 6). The FRET assay demonstrates ATP- and enzyme-dependent helicase activity and is linear in its response for at least 30 minutes at room temperature. It is tolerant of DMSO concentrations up to at least 2%.
Radiometric assays of helicase activity were performed as described previously31, 33 utilizing a radiolabeled sixty-mer:M13 partial duplex substrate (Table 6) possessing 5-nucleotide forks at both the 5′ and 3′ ends. A standard 20μl reaction volume contained 25 mM Tris-HCl, (pH 7.5), 10 mM MgCl2, 10% glycerol, 5 mM DTT and 0.1 mM ATP, 17 fmol (1-2 × 104 cpm/μl) of substrate and the indicated amount of helicase. The mixtures were incubated at 30°C for 15 min and the reactions were terminated by the addition of 4 μl of 2.5% SDS, 60 mM EDTA, and 1% bromophenol blue. Displacement was monitored by polyacrylamide gel electrophoresis, followed by autoradiography.
Two small molecule libraries were screened in this study. The Microbiotix, Inc. collection (MBX) was screened for inhibitors of B. anthracis helicase, and the collection at the National Screening Laboratory for the Regional Centers of Excellence in Biodefense and Emerging Infectious Disease (NSRB) at Harvard Medical School (Boston, MA) was screened for inhibitors of S. aureus helicase. Both libraries were comprised of compounds purchased from commercial vendors, and while the overlap between the two libraries has not been calculated, it is likely to be significant because some of the same vendors were used. Compounds in the Microbiotix screening deck were purchased from Chembridge (San Diego, CA), Timtec (Newark, DE), and ChemDiv (San Diego, CA) while the NSRB collection contained compounds from Asinex (Moscow), Enamine (Kiev, Ukraine), Life Chemicals, Inc. (Burlington, Ontario, CA), and Maybridge (Fisher Scientific International), as well as Chembridge and ChemDiv. Compounds were selected in the molecular size range of 200 to about 500 Da and were filtered to remove unwanted and known cytotoxic fragments, including metal complexes, highly conjugated ring systems, oxime esters, nitroso and strong Michael acceptors. Compounds in the Microbiotix collection were diluted in 96-well master plates at 2.5 mM in DMSO at -20°C. Master plates were thawed at room temperature on the day of the screen, and 1 μl of compound was added to the 384-well screening plates by means of a Sciclone ALH 3000 liquid handling robot (Caliper, Inc.) and a Twister II Microplate Handler (Caliper, Inc.). Compounds in the NSRB collection were added to assay plates by pin transfer. The first two columns of wells contained the complete reaction without inhibitors (postive control), and the last two columns contained the complete reaction, but no enzyme (negative control). Screening wells contained 30 μl volume consisting of ~80 μM compound, 10 nM annealed oligo duplex, 15× capture strand, 63 ng helicase (12 nM monomer), 30 mM Tris-HCl (pH 7.9), 0.01% Triton X-100, 0.5 mM DTT, 2 mM MgCl2, and 25 mM NaCl. The reaction was initiated by the addition of ATP to 2.5 mM final concentration using a Wellmate automated microplate dispenser. Plates were incubated at room temperature for 30 min, and fluorescence was read at 530 nm in an Envision 2102 Multilabel HTS Counter (Perkin Elmer) with excitation at 490 nm. Strand unwinding catalyzed by helicase resulted in loss of quenching and a linear increase in fluorescence during the 30 minute incubation at room temperature. The capacity of the screens was 50,000 compounds in singlet per week (7,000 compounds per day). The plate Z' value 34 averaged 0.59, and the signal:background ratio averaged 3.6. All screening data, including the z-score, and confirmation and validation data were stored in one central database (CambridgeSoft's ChemOffice 11.0). Validated hits were re-ordered from the vendor and confirmed to be >95% pure and to be of the expected mass by LC-MS analysis. Compounds for SAR analysis were ordered from Chembridge, Inc.
Hits were tested for inhibition of purified P. aeruginosa AmpC β-lactamase in the presence of various concentrations of Triton X-100 to detect compounds acting promiscuously by a colloidal aggregate mechanism as described previously.11 The ability of compounds to inhibit B. subtilis Pol IIIC DNA polymerase was assayed by using a 96-well plate format version of the standard DNA polymerase assay as described previously.35 Note that since the substrate for the Pol IIIC assay is DNase-activated chromosomal DNA, no helicase function is required for polymerization.35 Inhibition of B. subtilis DNA gyrase by compounds was measured by a published gel mobility assay.36 Inhibition of the ssDNA-stimulated ATPase activity of B. anthracis helicase was measured (10 μl volume, 100 μM inhibitor) in the presence of 0.5 mM ATP and 200 pmole of M13mp19 single-stranded DNA as previously described.29
Values for Ki and IC50 for inhibitory compounds were determined by using the FRET-based assay under the same conditions as described for screening except that annealed oligonucleotide substrate or inhibitor concentrations were varied. All IC50 values were determined in duplicate using a 10-point curve consisting of two-fold dilutions of inhibitory compound from 100 μM to 0.2 μM. Substrate and inhibitor concentrations for kinetic experiments are noted in the figure. Rapid assessments of mode of inhibition were done by the method of Wei et al.37 by determining the variation in IC50 values over a range of substrate and inhibitor concentrations.
Hits were examined in a microplate ethidium bromide displacement assay10 to determine whether compounds bind to the DNA duplex substrate.
Compounds were tested for inhibition of ATP-dependent DNA replication in permeabilized ΔpolA B. anthracis Sterne cells assay according to a modification of the method of Brown et al.38 This assay was used to confirm whether the compounds are capable of inhibiting an intact bacterial replisome.
MICs were determined by the broth microdilution method described in the CLSI (formerly NCCLS) guidelines.39 All MICs were expressed in μM to facilitate comparisons with IC50 and CC50 values and were determined in duplicate using a 10-point curve consisting of two-fold dilutions of inhibitory compound from 100 μM to 0.2 μM. For bactericidal tests, inhibitors were examined in a standard method of broth culture of B. anthracis Sterne cells followed by plating on LB agar media and counting colony-forming units.40
The cytotoxic concentration (CC50) of compounds versus cultured mammalian cells (HeLa) was determined as the concentration of compound that inhibits 50% of the conversion of MTS to formazan.41 Values were determined in duplicate using a 10-point curve consisting of two-fold dilutions of inhibitory compound from 100 μM to 0.2 μM. The “selectivity index” of a given agent is defined as the ratio of its mammalian cell cytotoxicity to its MIC value against B. anthracis (i.e. CC50/MIC).
A poriton of the high-throughput screening capability reported here was provided by the National Screening Laboratory for the Regional Centers of Excellence in Biodefense and Emerging Infectious Diseases (NIAID U54 AI057159). We thank Dr. Su Chiang (NSRB, Harvard Medical School, Boston, MA) for assistance and coordination of the screening of the NSRB library. We thank Dr. Michelle Butler and Ms. Lauren Kustigian for MIC data. This work was supported by SBIR Grants AI063712 and AI064974 from NIAID/NIH.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.