Inhibition of Yersinia pestis DNA adenine methyltransferase in vitro by a stibonic acid compound: identification of a potential novel class of antimicrobial agents

Abstract

Background and purpose: Multiple antibiotic resistant strains of plague are emerging, driving a need for the development of novel antibiotics effective against Yersinia pestis. DNA adenine methylation regulates numerous fundamental processes in bacteria and alteration of DNA adenine methlytransferase (Dam) expression is attenuating for several pathogens, including Y. pestis. The lack of a functionally similar enzyme in humans makes Dam a suitable target for development of novel therapeutics for plague.

Experimental approach: Compounds were evaluated for their ability to inhibit Dam activity in a high-throughput screening assay. DNA was isolated from Yersinia grown in the presence of lead compounds and restricted to determine the effect of inhibitors on DNA methylation. Transcriptional analysis was undertaken to determine the effect of an active inhibitor on virulence-associated phenotypes.

Key results: We have identified a series of aryl stibonic acids which inhibit Dam in vitro. The most active, 4-stibonobenzenesulfonic acid, exhibited a competitive mode of inhibition with respect to DNA and a K(i) of 6.46 nM. One compound was found to inhibit DNA methylation in cultured Y. pestis. The effects of this inhibition on the physiology of the cell were widespread, and included altered expression of known virulence traits, including iron acquisition and Type III secretion.

Conclusions and implications: We have identified a novel class of potent Dam inhibitors. Treatment of bacterial cell cultures with these inhibitors resulted in a decrease in DNA methylation. Expression of virulence factors was affected, suggesting these inhibitors may attenuate bacterial infectivity and function as antibiotics.

Figure 1

Break light Dam activity assay and raw assay data. (A) Hemimethylated break light oligonucleotide 1 containing a GATC methylation site is fully methylated in a reaction catalyzed by Dam forming oligonucleotide 2, which is a substrate for the restriction enzyme DpnI. Cleavage by DpnI results in the separation of fluorophore (fluorescein) and quencher (dabcyl) and a subsequent proportional increase in fluorescence. (B) Fluorescence signal against time for positive (containing no library compound) and negative (containing no AdoMet and no library compound) control assays. (C) Fluorescence signal against time for full assay supplemented with inhibitor compound 3 (1, 10 or 100 μM).

Figure 2

Structure of NCI/DTP library screen hits. Structures of NSC compounds: 10460 (1), 93739 (2), 13776 (3), 88915 (4), 109268 (5) and 166687 (6).

Figure 3

Results of screening and counter-screening for NCI/DTP library screen hits. Bar chart showing inhibition of Yersinia pestis Dam, inhibition of DpnI and percentage displacement of thiazole orange from DNA. All percentages shown are relative to positive control assays containing no library compound.

Figure 4

Arylstibonic acid sub-library and IC₅₀ plots of arylstibonic acid inhibitors. (A) Structure of the arylstibonic acid sub-library (individual compound structures are shown in Supplementary Section S7). (B) IC₅₀ plots for compounds 3 and 13 showing increase in potency, data fitted to equation 5 to give the results shown in Table 3. Structure of compound 3 is shown; for compound 13, R₁ and R₂ = H and R₃ = SO₃H.

Figure 5

Inhibition of Y. pestis Dam by compounds 3 and 13. (A) Varying DNA concentration and (B) varying AdoMet concentration. Each substrate was tested against a range of concentrations of inhibitor 3 (0, 0.77, 1.92, 4.80, 12.00, 30 and 75 μM). (C) Varying DNA concentration and (D) varying AdoMet concentration. Each substrate was tested against a range of concentrations of inhibitor 13 (0, 6.54, 15.36, 38.40, 96, 240 and 600 nM). For the experiments in which the concentration of oligonucleotides 1 was varied, data were fitted to a competitive inhibition model to give K_i values of 4.90 ± 1.78 nM for compound 13 and 0.70 ± 0.18 μM for compound 3. A non-competitive model was used for experiments in which AdoMet was varied and gave K_i values of 47.80 ± 3.02 nM for compound 13 and 4.52 ± 0.23 μM for compound 3.

Figure 6

Inhibition of Dam activity by compound 14 in (A) Y. pseudotuberculosis and (B) Y. pestis. Genomic DNA (2 μg) was isolated using the Puregene DNA isolation kit (Gentra Systems, Minneapolis, USA) from (A) Y. pseudotuberculosis and (B) Y. pestis. The DNA was digested by one of the restriction endonucleases, MboI, Sau3AI or DpnI then analysed by electrophoresis on 0.7% (w/v) agarose gels. Uncut DNA appears as a dark smear high on the gel. Digestion results in degradation of the DNA to a range of small fragments, which make a faint smear. In (A), Lane 1: Markers (DriGestIII, Amersham), Lanes 2-5: DNA isolated from Y. pseudotuberculosis grown in the presence of compound 19 either uncut (Lane 2) or after incubation with DpnI (Lane 3), MboI (Lane 4) or Sau3AI (Lane 5). Lanes 6-9: DNA isolated from Y. pseudotuberculosis grown in the presence of compound 14 either uncut (Lane 6) or after incubation with DpnI (Lane 7), MboI (Lane 8) or Sau3AI (Lane 9). Uncut DNA is visible in Lanes 2, 4 and 6, but not in the remaining lanes, showing that the DNA has been digested in these samples to small fragments. In (B). Lanes 1-4: DNA isolated from Y. pestis grown in the presence of compound 14 either uncut (Lane 1) or after incubation with DpnI (Lane 2), MboI (Lane 3) or Sau3AI (Lane 4). Lane 5: Markers (DriGestIII, Amersham). Uncut DNA is visible in Lane 1, as expected. DNA is not visible in Lanes 2 and 4, showing that the DNA has been digested in these samples to small fragments. A smear is visible in Lane 3, showing that the DNA has been digested, but the fragments are larger than was generated in Lanes 2 and 4.

Figure 7

Growth curves for Y. pestis GB in BAB, BAB supplemented with DMSO or BAB supplemented with DMSO and compound 14. Data fitted to a sigmoid of the form f = a/{1 + exp[−(x − x₀)/b]}.