Optimization of a novel potent and selective bacterial DNA helicase inhibitor scaffold from a high throughput screening hit

Abstract

Benzobisthiazole derivatives were identified as novel helicase inhibitors through high throughput screening against purified S. aureus (Sa) and B. anthracis (Ba) replicative helicases. Chemical optimization has produced compound 59 with nanomolar potency against the DNA duplex strand unwinding activities of both B. anthracis and S. aureus helicases. Selectivity index (SI = CC₅₀/IC₅₀) values for 59 were greater than 500. Kinetic studies demonstrated that the benzobisthiazole-based bacterial helicase inhibitors act competitively with the DNA substrate. Therefore, benzobisthiazole helicase inhibitors represent a promising new scaffold for evaluation as antibacterial agents.

The increasing prevalence of antibiotic resistant strains of bacterial pathogens presents a major unmet medical need. For example, the incidence of skin and soft tissue infections with community acquired methicillin-resistant Staphylococcus aureus (MRSA) has increased over five-fold since 2003.¹ In addition, a recent report revealed that 28% of enterococci cultured from 25 North American intensive care units (ICUs) were resistant to vancomycin (VRE).² The development of new antibiotics against underexploited targets with novel mechanisms of action is a vital part of the solution to these problems because such antibiotics will not be affected by preexisting target based resistance mechanisms. The replicative DNA helicase is an essential component of the DNA replication pathway, acting early and catalyzing a rate limiting step in replication, but it is currently untargeted by antibacterial agents.

Replicative DNA helicase is a member of a drug-validated pathway and, along with gyrase, topoisomerase IV, and DNA polymerase III, is essential to bacteria.^3–7 The primary structures of bacterial replicative helicases differ significantly from those of their eukaryotic and human counterparts,^8,9 indicating that bacteria-specific inhibitors of helicase may be developed. These features make it particularly attractive as a target for the discovery of new antibacterial therapeutics.

The replicative DNA helicases from E. coli, S. aureus, and P. aeruginosa have been targeted previously in anti-infective screens,^10–17 but few hits have been described, and none have progressed further in drug development due to poor potency and inadequate selectivity. Two distinct X-ray crystal structures have been reported: one shows a hexameric DnaB helicase in complex with a helicase binding fragment of primase,¹⁸ and another shows that the DnaB hexamer adopts a closed spiral staircase quaternary structure in complex with ATP mimic GDP-AlF4 and ssDNA.¹⁹ The two structures suggest that helicase may exist in both inactivated and activated forms during the bacterial DNA replication process. Structure-based approaches to target both the inactivated and activated forms of DnaB helicase may aid in the discovery of novel bacterial DNA helicase inhibitors.

We have previously discovered a coumarin-based DNA helicase inhibitor series through a high throughput screening campaign, and chemical optimization yielded compounds with antibacterial activities against several Gram-positive species including multiple clinically relevant ciprofloxacin-resistant MRSA strains.^20,21 Herein we report chemical optimization and biological evaluation of a novel series of DNA bacterial helicase inhibitors based on a benzobisthiazole scaffold.

Benzobisthiazole derivatives were identified as novel inhibitors through high throughput screening against S. aureus (Sa) and B. anthracis (Ba) helicases. The screening hit compound 1 (Fig. 1) demonstrated antihelicase activities (IC₅₀ (Ba) = 5 μM and IC₅₀ (Sa) = 12 μM) and minimal cytotoxicity (CC₅₀ > 100 μM). Initial investigation of structure-activity relationships (SARs) focused on the benzamide portion of the benzobisthiazole scaffold. Thirty-two analogs (compounds 4–35) with various substitutions on the phenyl ring of the benzamide portion and twelve analogs (compounds 36–47) with phenyl ring replacements were purchased from an outside vendor (Life Chemicals, Inc.). Their biological activities were evaluated in a fluorescence resonance transfer (FRET)-based helicase strand unwinding assay²⁰ to measure concentration-dependent inhibition of Bacillus anthracis DNA replicative helicase, and the results are summarized in Tables 1 and 2.

Figure 1.

The structure of HTS hit 1.

Table 1.

Bacillus anthracis helicase inhibition by benzobisthiazole compounds 1–33.

Compd.	R	IC50 Ba Helicase (μM)
1	3-N[-(C=O)CH2CH2(C=O)-]	5.0
2	4-N(CH2CH3)2	>100
3	4-tBu	>100
4	4-OPh	>100
5	4-OiPr	>100
6	4-SO2N(CH2CH3)2	>100
7	4-CH3	7.6
8	4-CN	11.0
9	4-OCH2CH3	6.9
10	4-Cl	5.0
11	4-COOCH3	3.7
12	4-N(CH3)2	2.4
13	4-OCH3	2.2
14	4-N[-(C=O)CH2CH2(C=O)-]	1.3
15	3-N(CH3)2	36.0
16	3-OCH3	12.5
17	3-Cl	5.3
18	3-F	3.3
19	3-NO2	2.8
20	3-Br	2.3
21	2-OPh	>100
22	2-Cl	>100
23	2-Br	>100
24	2-F	~50
25	2-CH3	6.6
26	2,4-(OCH3)2	>100
27	2,6-(OCH3)2	>100
28	2,6-F2	64.5
29	3,4-(CH3)2	3.2
30	3,4-(OCH2CH2O)	3.0
31	3,4-(OCH3)2	2.1
32	3,5-(OCH3)2	1.7
33	3,4,5-(OCH3)3	0.7

Table 2.

Bacillus anthracis and Staphyloccocus aureus helicase inhibition by benzobisthiazole compounds 34–45.

Compd.	R1	IC50 Ba Helicase (μM)
34	cyclopropyl	>100
35	CH2Ph	47
36	CH2CH2Ph	>100
37		>100
38		>100
39		>50
40		>100
41	2-furan	102
42	2-thiophene	90
43		>50
44		>50
45	2-pyrazine	28

Substituents on the phenyl ring of the benzamide portion dramatically affected the antihelicase activity of the benzobisthiazoles (Table 1). In general, compounds with bulky substituents (compounds 2–6) were inactive vs B. anthracis helicase, while smaller substituents, such as F, Cl, Br, CN, CH₃, CO₂CH₃, OCH₃, and OCH₂CH₃ were tolerated at the 3- or 4-positions (compounds 7–20). Substituents at the 2-position of the phenyl ring were not tolerated except for the 2-CH₃ group (compound 25). Disubstitution at the 3,4- or 3,5-positions with CH₃ or OCH₃ groups on the phenyl ring was tolerated. For example, compounds 29–32 with substituents 3,4-(CH₃)₂, 3,4-(OCH₂CH₂O), 3,4-(OCH₃)₂, and 3,5-(OCH₃)₂ displayed 1.7–3.2 μM IC₅₀ values vs B. anthracis helicase, while compounds with disubstitution at the 2,4- or 2,6-positions (26–28) exhibited weak or no inhibitory activity. Compound 33, with 3,4,5-(OCH₃)₃ substitution on the phenyl ring, showed the best potency with an IC₅₀ value of 0.7 μM in this initial investigation of probing the substitution effect on the antihelicase activity.

The effect of replacement of the phenyl ring with various groups was also investigated in the B. anthracis DNA helicase assay, and the results are shown in Table 2. Replacement of the phenyl ring with alkyl, arylalkyl, naphthyl or heteroaryl groups (compounds 34–44) significantly decreased potency, except for compound 45 with a pyrazine replacement, which exhibited modest activity (IC₅₀ = 28 μM).

The most active B. anthracis helicase inhibitor, compound 33, also exhibited potent inhibitory activity vs S. aureus DNA helicase (IC₅₀ = 0.4 μM) without detectable cytotoxicty (CC₅₀ >100 μM), while compound 16, which bears a 3-OCH₃ group on the phenyl ring, inhibited S. aureus DNA helicase with an IC₅₀ value of 6.6 μM. To evaluate the SARs on the methylthio side of the benzobisthiazole core structure, we synthesized a series of analogs of two precursors 33 and 16, by further transforming the methylthio group to various amines, and the synthesis is shown in Scheme 1.

Scheme 1.

Reagents and conditions: a, benzoyl chlorides, CH₂Cl₂, yields 31–95%; b, aq. KMnO₄, glacial HOAc, dioxane, 73% for 47, and 44% for 48; c, R²R³NH, DMF, 80–100 °C, yields 20–83%.

Commercially available aminobenzobisthiazole compound 46 was treated with corresponding benzoyl chlorides, to yield amides 16 and 33.²² Treatment of compounds 16 and 33 with KMnO₄ under acidic conditions produced sulfones 47 and 48, respectively, which were treated with amines to produce benzobisthiazole analogs 49–63.^{23, 24} To evaluate whether the amide bond was required for antihelicase activity, amino compound 46 was converted to benzylamino analog 64 through reductive amination, and an imine analog 65 was also synthesized through a condensation reaction (Scheme 2).²⁵ The biological activities of these analogs were evaluated vs both Bacillus anthracis and S. aureus DNA replicative helicases, and cytotoxicities were measured in a serum free MTT assay²¹. Results are summarized in Table 3.

Scheme 2.

Reagents and conditions: a, 3,4,5-trimethoxybenzaldehyde, Na(OAc)₃BH, glacial HOAc, DMSO, r.t., 48%; b, 3,4,5-trimethoxybenzaldehyde, toluene, reflux, 67%.

Table 3.

Bacillus anthracis and Staphyloccocus aureus helicase inhibition by benzobisthiazole compounds 16, 33 and 49–65.

Compd.	R1	R2	IC50 B. anthracis Helicase (μM)	IC50 S. aureus Helicase (μM)	HeLa CC50 in SFM (μM)
16	S-CH3	3-OCH3	12.5	6.6	12.4
49	NH-CH3	3-OCH3	0.5	2.2	>100
50	NH-propyl	3-OCH3	1.1	1.3	<0.08; 15.6
51	NH-isopropyl	3-OCH3	65.0	80	n.d.
52	NH-isobutyl	3-OCH3	13.0	9.0	44
53	NH-cyclohexyl	3-OCH3	90	>100	n.d.
54	NH-Bn	3-OCH3	75	>100	n.d.
55	NHCH2CH2N(CH3)2	3-OCH3	1.0	1.0
56		3-OCH3	0.5	2.2	9.6
57		3-OCH3	2.3	3.0	>100
58	NHCH2CH2OH	3-OCH3	2.2	4.2	>100
33	S-CH3	3,4,5-(OCH3)3	0.7	0.4	>100
59	NH-CH3	3,4,5-(OCH3)3	0.2	0.2	>100
60	NHCH2CH2N(CH3)2	3,4,5-(OCH3)3	0.2	0.2	<0.78
61		3,4,5-(OCH3)3	0.2	0.2	5.7
62		3,4,5-(OCH3)3	0.2	0.2	2.8
63	NHCH2CH2OH	3,4,5-(OCH3)3	0.2	0.3	>100
64			>100	>100	n.d.
65			>100	>100	n.d.

Compared to the methylthio compound 16, methylamino analog 49 exhibited improved biological activity vs both Ba and Sa DNA helicases, and its cytotoxicity (CC₅₀) was greater than 100 μM. Increasing the size of the alkylamino group revealed that propylamino was tolerated; however, isopropylamino, cyclohexylamino and benzylamino compounds (51, 53, 54) showed significantly reduced activity, while the isobutylamino compound 52 displayed modest inhibition vs both of the helicases. Interestingly, compounds 55–58, which bear additional polar groups, such as N(CH₃)₂, morpholino, or OH groups through a linker attaching to the aminobenzobisthiazole core, exhibited improved potency with IC₅₀ values ranging between 1.0–2.3 μM and 1.0–4.2 μM vs Ba and Sa helicases, respectively. This finding suggests that the polar groups make additional, favorable interactions with the helicase enzymes.

Similarly, compared to the methylthio compound 33, methylamino analog 59 also displayed improved helicase inhibitory activity (Ba helicase IC₅₀ = 0.2 μM; Sa helicase IC₅₀ = 0.2 μM) without detectable cytotoxicity (CC₅₀ >100 μM), and selectivity indices (CC₅₀/IC₅₀) were more than 500. Analogs 60–63 all inhibited Ba and Sa helicases with IC₅₀ values between 0.2–0.3 μM. Furthermore, compound 63 did not show cytotoxicity (CC₅₀ >100 μM). Compounds 64 and 65, in which the amide linker was replaced with -NHCH₂- and -N=CH- linkers, lost activities vs both Ba and Sa helicases, suggesting that the amide bond is required for inhibitor binding, probably because of favorable hydrogen bonding interactions.

To determine the mode of inhibition for the benzobisthiazole scaffold, a dilution series of compound 49 was added to the helicase reaction in the presence of varying concentrations of the two substrates, ATP and oligonucleotide. The IC₅₀ value of compound 49 varied less than 2-fold (2.0 to 3.8 μM) in response to changes in the ATP concentration (Fig. 2A), but varied considerably (>6.5-fold) when oligonucleotide concentrations were altered (Fig. 2B). These results suggest that compound 49 is noncompetitive with ATP but competitive with the DNA substrate.²⁶ Indeed, a Dixon plot (Fig. 2C) of the DNA variation results is consistent with competitive inhibition with a Ki value of 2 μM, which is in good agreement with the IC₅₀ value at the lowest oligonucleotide concentration tested (3 nM). Similar results were also obtained for compounds 59 and 63 in kinetic studies (data not shown). These results indicate that helicase inhibitors with the benzobisthiazole scaffold act with a different mechanism of action than the previously reported coumarin-based helicase inhibitors, which are noncompetitive with both ATP and DNA.²¹ However, neither the coumarin nor the benzobisthiazole inhibitors have any significant effect on the single-strand stimulated ATPase activity of S. aureus helicase at concentrations over 100-fold higher than their IC₅₀ values (Fig. 2D).²¹ X-ray crystallography studies have revealed two distinct hexameric DnaB helicase structures: one showed a flat structure without DNA and NTP association, and the other formed a closed spiral staircase quaternary structure in complex with DNA and ATP mimic GDP-AIF4, suggesting that the former structure represents an inactivated form of DnaB helicase, and the latter one represents an activated form of DnaB helicase. The coumarin type helicase inhibitors are kinetically noncompetitive with both DNA and ATP substrates, suggesting that the coumarin-based inhibitors are likely bound to the inactivated form of the helicase, while the benzobisthiazole derivatives exhibit competitive kinetics with the DNA substrate, suggesting that the benzobisthiazole-based compounds may inhibit the activated form of the bacterial helicase. Structural information for the inhibitor-DnaB helicase complex is needed to further elucidate the mechanism of action of these small molecule helicase inhibitors and provide guidance for the discovery of novel bacterial helicase inhibitors.

Figure 2.

Mode of inhibition of benzobisthiazole 49. A dilution series of compound 49 was examined in the FRET-based S. aureus helicase assay²⁰ with varying concentrations of ATP (A) or DNA (B) to determine IC₅₀ values. Data were plotted in GraphPad Prism 5.0 using 4-parameter curve fitting. (C) The results from varying DNA are also shown in a Dixon plot (1/V vs [I]). (D) Compounds 49 and 59 failed to exhibit significant inhibition of the single-stranded DNA stimulated ATPase activity of S. aureus helicase as detected by malachite green.²¹

In addition to potently inhibiting replicative helicase, many benzobisthiazole analogs also inhibit the growth of bacterial cells (data not shown). However, the growth inhibition manifests primarily as a reproducible lengthened lag phase, delaying exponential growth, and is not sufficient to produce MIC values, possibly due to poor bacterial cellular permeability and/or short half-life of the parent compounds in cells. Nonetheless, incorporation of radiolabeled precursors into DNA in permeabilized preparations of B. anthracis cells ²⁷ was reduced significantly by compounds 62 and 63 with IC₅₀ values of 10 and 17 μM, respectively (Fig. S1). These results indicate that the benzobisthiazoles inhibit DNA synthesis in bacterial cells, and such inhibition could be responsible for the modest growth inhibition observed in live cells.

In conclusion, we have identified a novel class of potent and selective benzobisthiazole-based bacterial DnaB helicase inhibitors through high throughput screening. Subsequent structure-activity relationship analysis and chemical optimization led to substantial improvement of biological activity and identification of nanomolar compounds with selectivity indices of more than 500. To the best of our knowledge, these compounds are the most potent bacterial replicative helicase inhibitors reported to date. Further optimization of the benzobisthiazole-based helicase inhibitors may provide a novel small molecule drug for antibacterial therapy.