Plumbagin inhibits cytokinesis in Bacillus subtilis by inhibiting FtsZ assembly – a mechanistic study of its antibacterial activity

Authors

  • Anusri Bhattacharya,

    1. Department of Chemistry, Indian Institute of Technology Bombay, Mumbai, India
    2. Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Mumbai, India
    Search for more papers by this author
  • Bhavya Jindal,

    1. Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Mumbai, India
    Search for more papers by this author
  • Parminder Singh,

    1. Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Mumbai, India
    Search for more papers by this author
  • Anindya Datta,

    1. Department of Chemistry, Indian Institute of Technology Bombay, Mumbai, India
    Search for more papers by this author
  • Dulal Panda

    Corresponding author
    1. Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Mumbai, India
    • Correspondence

      D. Panda, Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Mumbai 400076, India

      Fax: +91 22 2572 3480

      Tel: +91 22 2576 7838

      E-mail: panda@iitb.ac.in

    Search for more papers by this author

Abstract

The assembly of FtsZ plays a central role in construction of the cytokinetic Z-ring that orchestrates bacterial cell division. A naturally occurring naphthoquinone, plumbagin, is known to exhibit antibacterial properties against several types of bacteria. In this study, plumbagin was found to perturb formation of the Z-ring in Bacillus subtilis 168 cells and to cause elongation of these cells without an apparent effect on nucleoid segregation, indicating that it may inhibit FtsZ assembly. Furthermore, it bound to purified B. subtilis FtsZ (BsFtsZ) with a dissociation constant of 20.7 ± 5.6 μm, and inhibited the assembly and GTPase activity of BsFtsZ in vitro. Interestingly, plumbagin did not inhibit either the assembly or GTPase activity of Escherichia coli FtsZ (EcFtsZ) in vitro. Using docking analysis, a putative plumbagin-binding site on BsFtsZ was identified, and the analysis indicated that hydrophobic interactions and hydrogen bonds predominate. Based on the in silico analysis, two variants of BsFtsZ, namely D199A and V307R, were constructed to explore the binding interaction of plumbagin and BsFtsZ. The effects of plumbagin on the assembly and GTPase activity of the variant BsFtsZ proteins in vitro indicated that the residues D199 and V307 may be involved in the binding of plumbagin to BsFtsZ. The results suggest that plumbagin inhibits bacterial proliferation by inhibiting the assembly of FtsZ, and provide insight into the binding site of plumbagin on BsFtsZ, which may help in the design of potent FtsZ-targeted antibacterial agents.

Abbreviations
ANS

8-anilino-1-naphthalenesulfonic acid

BsFtsZ

Bacillus subtilis FtsZ

EcFtsZ

Escherichia coli FtsZ

TNP-GTP

2′,3′-O-(2,4,6-trinitrocyclohexadienylidine)-GTP

Introduction

The assembly and disassembly of the bacterial cell division protein FtsZ play an important role in bacterial cytokinesis [1-5]. At the onset of the bacterial cell division process, FtsZ migrates to the division site where it polymerizes to form a highly dynamic polymeric structure termed the Z-ring [6, 7], and recruits several accessory proteins to form the divisome [2]. FtsZ is a tubulin-like GTPase, and hydrolysis of GTP is thought to regulate the dynamics of FtsZ assembly [8]. Inhibition of FtsZ assembly has been shown to cause bacterial cell elongation and to inhibit bacterial proliferation [3, 9, 10]. In addition to its critical role in bacterial cytokinesis, FtsZ is a highly conserved protein in prokaryotes [11], thereby making it a promising antibacterial drug target [12-15].

Plumbagin (5-hydroxy-2-methyl-1,4-naphthoquinone) is a secondary plant metabolite exhibiting several biological activities, including its ability to inhibit proliferation of mammalian, fungal and bacterial cells [16-20]. Plumbagin has been shown to have toxicity against several pathogenic and non-pathogenic microbes [21-23]. For example, it has been shown to inhibit the proliferation of various bacteria such as Staphylococcus aureus, Pseudomonas aeruginosa, Bacillus subtilis, Proteus vulgaris, Mycobacterium smegmatis mc2 155 and Mycobacterium tuberculosis H37Rv [22, 23]. However, the antimicrobial properties of plumbagin are restricted to a few species of bacteria, as Escherichia coli and Salmonella typhimurium have been shown to grow efficiently in its presence [22].

The present study demonstrated that plumbagin inhibited the proliferation of B. subtilis 168 and M. smegmatis cells. In addition, plumbagin treatment disrupted the formation of functional Z-rings and increased the length of B. subtilis 168 cells. Several agents are known to induce cell elongation and to inhibit bacterial cell division by perturbing FtsZ assembly [24-28]. Recently, plumbagin has been shown to bind to tubulin and to inhibit tubulin assembly in vitro [29]. As FtsZ is a prokaryotic homolog of tubulin, and plumbagin caused elongation of bacterial cells, we wished to examine the effect of plumbagin on FtsZ assembly. We found that plumbagin bound to FtsZ, and inhibited the assembly and GTPase activity of BsFtsZ in vitro. However, it had no effect on the assembly and GTPase activity of E. coli FtsZ (EcFtsZ), suggesting that FtsZ from different bacteria may have different ligand-binding properties.

In this study, we provide evidence suggesting that plumbagin inhibits bacterial proliferation by inhibiting FtsZ assembly. In addition, using docking analysis and site-directed mutagenesis, we have characterized the binding site of plumbagin on BsFtsZ. Two residues in BsFtsZ (aspartic acid at position 199 and valine at position 307) were found to be important for the binding interactions of plumbagin and BsFtsZ. The study provides a rationale for developing potent plumbagin analogs.

Results

Plumbagin inhibits the proliferation of B. subtilis 168 and M. smegmatis cells but does not affect E. coli K12 cells

Using the standard agar dilution method [28, 30], the minimum inhibitory concentrations of plumbagin for B. subtilis 168 and M. smegmatis cells were found to be 29 and 31 μm, respectively. Consistent with a previous study [22], proliferation of E. coli K12 cells was found to be unaffected in the presence of plumbagin.

Plumbagin-induces filamentation in B. subtilis 168 cells

Plumbagin inhibited the growth rate of B. subtilis 168 cells in a concentration-dependent manner (Fig. 1A). Furthermore, plumbagin-treated B. subtilis 168 cells were found to be elongated several-fold compared to untreated cells (Fig. 1B). For example, upon incubation with 5 μm plumbagin for 4 h, the mean cell length increased sevenfold from 4 ± 0.1 to 28 ± 2 μm. Moreover, the cell length distribution of vehicle-treated and plumbagin-treated cells indicates that plumbagin blocks cytokinesis in bacteria (Fig. 1C).

Figure 1.

Plumbagin inhibits the proliferation of B. subtilis 168 cells and induces filamentation in these cells. (A) Effect of plumbagin on the growth of B. subtilis 168 cells. B. subtilis 168 cells were grown in the absence of plumbagin (closed square) or the presence of 2 μm plumbagin (open circle), 5 μm plumbagin (closed upright triangle), 10 μm plumbagin (open downward triangle) and 15 μm plumbagin (closed leftward-pointing triangle) for 6 h. (B) B. subtilis 168 cells were grown in the absence of plumbagin (i) or the presence of 5 μm plumbagin (ii) at 37 °C for 4 h, and the cell morphology was observed under microscope. Scale bar = 10 μm. (C) Plumbagin increased the length of B. subtilis 168 cells.

Plumbagin inhibits Z-ring formation

Plumbagin inhibited cytokinesis in B. subtilis 168 cells, indicating that it may perturb Z-ring formation in bacteria. As expected, plumbagin treatment was found to inhibit formation of the Z-ring in Bsubtilis 168 cells (Fig. 2). In addition, the number of Z-rings at the mid-cell position and the frequency of Z-rings per micrometer of the cell length were considerably reduced in the presence of plumbagin (Table 1). In the absence of plumbagin (control), 21, 63 and 10% of Bsubtilis 168 cells were found to contain one, two and four nucleoids, respectively. In the presence of 5 and 15 μm plumbagin, the percentage of cells containing one nucleoid was significantly reduced, and the majority of the cells contained more than two nucleoids (Table 1). However, plumbagin treatment did not affect the frequency of nucleoids per micrometer of the cell length, indicating that plumbagin does not perturb nucleoid segregation (Table 1).

Table 1. Effects of plumbagin on the Z-ring and nucleoids of B. subtilis 168 cells
DescriptionControlPlumbagin (5 μm)Plumbagin (15 μm)
Percentage of cells with Z-rings61%17%12%
Frequency of Z-rings per μm cell length0.14 ± 0.0070.03 ± 0.0040.02 ± 0.005
Frequency of nucleoids per μm cell length0.41 ± 0.030.38 ± 0.050.39 ± 0.06
Figure 2.

Plumbagin perturbs Z-ring formation in B. subtilis 168 cells. B. subtilis 168 cells were grown in the absence or presence of plumbagin (5 and 15 μm) at 37 °C for 4 h. Cells were fixed and stained using polyclonal FtsZ antibody (red). DNA was stained using DAPI (blue). Scale bar = 10 μm.

Plumbagin inhibits the assembly of BsFtsZ

Plumbagin disrupted Z-ring formation in B. subtilis 168 cells; therefore, we examined the effects of plumbagin on assembly and bundling of purified BsFtsZ in vitro. Plumbagin inhibited the polymerization of BsFtsZ in a concentration-dependent manner (Fig. 3A). For example, the extent of assembly was reduced by 26, 33 and 45% in the presence of 2, 5 and 10 μm plumbagin, respectively. In addition, the initial rates of increase in the light-scattering intensity of assembly of BsFtsZ in the absence and presence of 2, 5 and 10 μm plumbagin were 0.187, 0.143, 0.035 and 0.033 artibrary units per second, respectively indicating that plumbagin reduced the initial rate of assembly of BsFtsZ. In addition, the light-scattering traces revealed that plumbagin increased the lag phase of FtsZ assembly, indicating that plumbagin has an adverse effect on the nucleation rate of BsFtsZ. The effect of plumbagin on BsFtsZ assembly was further analyzed using transmission electron microscopy. Plumbagin was found to inhibit BsFtsZ assembly and to induce aggregation of BsFtsZ monomers at high concentrations (Fig. 3B).

Figure 3.

Plumbagin inhibits the assembly and GTPase activity of BsFtsZ. (A) BsFtsZ (8 μm) was polymerized in the absence of plumbagin (open square) or the presence of 2 μm plumbagin (closed circle), 5 μm plumbagin (open circle) or 10 μm plumbagin (closed downward triangle) at 37 °C in assembly buffer containing 25 mm PIPES (pH 6.8), 50 mm KCl, 10 mm MgCl2 and 1 mm GTP. (B) Electron micrographs of BsFtsZ polymers in the absence and presence of 20 μm plumbagin. Scale bars = 200 nm. (C) The amount of Pi released per mol BsFtsZ in the absence and presence of various concentrations of plumbagin was determined after 10 min of BsFtsZ assembly.

Plumbagin suppresses the GTPase activity of BsFtsZ

Plumbagin reduced the GTPase activity of BsFtsZ in vitro (Fig. S1). For example, the rate of GTPase activity of BsFtsZ was 2.5 and 1.7 mol Pi/mol FtsZ/min in the absence and presence of 10 μm plumbagin, respectively. Moreover, the effects of various concentrations of plumbagin on the GTPase activity of BsFtsZ were estimated after 10 min of FtsZ assembly (Fig. 3C). The results shown do not reflect true steady-state GTPase activities but the values that are influenced by both steady-state and initial GTPase activities. Plumbagin decreased the rate of GTP hydrolysis of BsFtsZ in a concentration-dependent fashion (Fig. 3C). For example, 24 μm plumbagin inhibited GTPase activity by 58%. The results indicate that plumbagin inhibits the rate of GTP hydrolysis of BsFtsZ either by disrupting binding of GTP to BsFtsZ or by modulating the assembly of BsFtsZ. A fluorescent analog of GTP, 2′,3′-O-(2,4,6-trinitrocyclohexadienylidine)-GTP (TNP-GTP), has been shown to bind to BsFtsZ with a Kd of 10 ± 2.7 μm [28]. While GTP was found to reduce the fluorescence of TNP-GTP in the presence of BsFtsZ, plumbagin did not reduce the fluorescence of TNP-GTP, suggesting that plumbagin did not compete with GTP for its binding site on BsFtsZ (Fig. S2).

Plumbagin interacts with BsFtsZ in vitro

The interaction of plumbagin with BsFtsZ was monitored using size-exclusion chromatography. BsFtsZ eluted in the void volume (~ 8.5 mL) while free plumbagin eluted in the column volume (~ 34 mL). However, when BsFtsZ was incubated with plumbagin and then loaded onto the same column, plumbagin co-eluted with BsFtsZ in the void volume, indicating that plumbagin was bound to BsFtsZ (Fig. 4A).

Figure 4.

Plumbagin binds to BsFtsZ in vitro. (A) BsFtsZ (15 μm) was incubated with plumbagin (60 μm) in 25 mm PIPES buffer, pH 6.8, and eluted through a G-25 column. Shown are the elution profiles of WT-BsFtsZ (open circle), plumbagin only (open square), plumbagin- (closed upright triangle) – WT-BsFtsZ complex (closed circle). (B) Effects of plumbagin on the tryptophan fluorescence of Y273W-BsFtsZ. Y273W-BsFtsZ (1 μm) was incubated without plumbagin (open circle) or with 1 μm plumbagin (closed upright triangle), 2 μm plumbagin (open square), 5 μm plumbagin (closed downward triangle), 10 μm plumbagin (open leftward-pointing triangle), 20 μm (closed rightward-pointing triangle), 30 μm plumbagin (closed diamond), 40 μm plumbagin (closed circle) or 50 μm plumbagin (closed square). (C) The changes in the fluorescence intensities of Y273W-BsFtsZ were plotted against plumbagin concentrations. The Kd of the binding interaction between Y273W-BsFtsZ and plumbagin was determined as described in 'Experimental procedures'. The experiment was performed five times. A representative curve from one such experiment is shown.

Intrinsic tryptophan fluorescence is a widely used tool to characterize binding of a ligand to a protein [24, 31]. As BsFtsZ does not have a tryptophan residue, we designed a variant (Y273W) in which a tryptophan residue replaces the tyrosine residue at position 273 [28]. Y273W-BsFtsZ exhibited characteristic tryptophan fluorescence when excited at 295 nm, and also displayed assembly characteristics similar to those of wild-type BsFtsZ (Fig. S3). Plumbagin quenched the tryptophan fluorescence of Y273W-BsFtsZ in a concentration-dependent manner (Fig. 4B). The dissociation constant of the interaction between Y273W-BsFtsZ and plumbagin was 20.7 ± 5.6 μm (Fig. 4C). Plumbagin did not alter the far-UV CD spectra of BsFtsZ, indicating that it did not perturb the secondary structure of BsFtsZ (Fig. S4).

Assembly of EcFtsZ is unaltered in the presence of plumbagin

Plumbagin did not inhibit the proliferation of E. coli K12 cells; therefore, we wished to monitor its effect on the assembly of EcFtsZ in vitro. The effects of plumbagin on the assembly kinetics of EcFtsZ were examined by 90° light scattering (Fig. S5A). Plumbagin had no detectable effect on assembly of EcFtsZ (Fig. S5A). In addition, we calculated the initial rates of polymerization of EcFtsZ in the absence and presence of 2–30 μm plumbagin for the first 100 s (shown as an inset to Fig. S5A). The initial rates did not reduce significantly with increasing concentrations of plumbagin. For example, the initial rates of polymerization of EcFtsZ were 4.1, 4.0, 3.1 and 3.5 artibrary units per second in the absence and presence of 10, 20 and 30 μm plumbagin, respectively. In addition, electron microscopic analysis of the assembly mixture suggested that plumbagin did not inhibit assembly of EcFtsZ (Fig. S5B). Furthermore, plumbagin had no effect on the GTPase activity of EcFtsZ after 10 min of assembly (Fig. S5C). The results suggest that plumbagin does not influence the assembly properties of EcFtsZ.

Putative binding site of plumbagin on BsFtsZ

Docking analysis indicated that the residues of BsFtsZ that constitute the plumbagin-binding site include R191, Q192, Q195, G196, D199, N263, T265, N299, V307 and T309 (residues within 4 Å). Hydrophobic and hydrogen-bonding interactions were found to exist between plumbagin and BsFtsZ. Four residues of BsFtsZ (R191, Q195, D199 and N299) were found to be hydrogen-bonded to plumbagin (Fig. 5). Two plumbagin-interacting residues of BsFtsZ (D199 and V307) were chosen for mutation to elucidate the binding site of plumbagin on BsFtsZ. The residues were selected because these residues either formed hydrogen bonds or were involved in hydrophobic interactions with plumbagin. In addition, these residues were found to differ from the corresponding residues of EcFtsZ. Moreover, docking analysis suggested that plumbagin may bind to EcFtsZ at a site that is different from that of BsFtsZ with an estimated Kd of 87 μm. The putative binding site of plumbagin was found to be at the N-terminal region of EcFtsZ. The residues within 4 Å of the binding site include G21, M104, T132, P134, E138, R142, N165, F182, A185, N186 and L189 (Fig. S6). This is interesting because plumbagin did not inhibit the polymerization and GTPase activity of EcFtsZ (Fig. S5). The interacting residues of BsFtsZ that were found to be similar to the corresponding EcFtsZ residues by sequence alignment were not selected for mutational studies. An alanine residue replaced the D199 residue, whereas V307 was replaced with a positively charged residue arginine. Two variants of BsFtsZ (D199A and V307R) were constructed, and the effects of plumbagin on the assembly properties of the variant proteins were analyzed to obtain insight into the binding site of plumbagin on BsFtsZ.

Figure 5.

Identification of the plumbagin-binding site on BsFtsZ. (A) Structure of plumbagin. (B) Plumbagin-binding site displayed on BsFtsZ. The BsFtsZ monomer is shown as a ribbon diagram, and the residues constituting the plumbagin-binding site are shown in sticks. (C) Detailed view of the residues of BsFtsZ interacting with plumbagin. Plumbagin is shown as green sticks and other sticks represent BsFtsZ residues. Green dashed lines show hydrogen bonds between the amino acids and plumbagin. The cut-off distance used for the hydrogen-bonding pairs is 3.6 Å between the donor and acceptor atoms. The figure was drawn using PyMol [51].

Plumbagin had no visible effect on assembly of V307R-BsFtsZ

Under the conditions used for assembly, V307R-BsFtsZ polymerized in a similar fashion to wild-type BsFtsZ (WT-BsFtsZ) and also displayed similar GTPase activity to that of WT-BsFtsZ, indicating that the assembly characteristics of the variant FtsZ are similar to those of native BsFtsZ. Plumbagin had no detectable effect on the assembly kinetics of V307R-BsFtsZ (Fig. 6A). In addition, electron microscopic analysis of the assembly milieu indicated that plumbagin did not inhibit assembly of the variant FtsZ (Fig. 6B). Furthermore, plumbagin did not inhibit the GTPase activity of V307R-BsFtsZ (Fig. 6C). The results together suggested that the V307R mutation in BsFtsZ abolished the inhibitory effect of plumbagin on assembly of BsFtsZ in vitro.

Figure 6.

Plumbagin has no detectable effect on the assembly and GTPase activity of V307R-BsFtsZ in vitro. (A) V307R-BsFtsZ (8 μm) in 25 mm PIPES (pH 6.8), 50 mm KCl, 10 mm MgCl2 and 1 mm GTP was polymerized at 37 °C in the absence (closed square) and presence of 2 μm plumbagin (closed circle), 5 μm plumbagin (closed upright triangle), 10 μm plumbagin (closed downward triangle), 20 μm plumbagin (closed leftward-pointing triangle) and 30 μm plumbagin (closed rightward-pointing triangle). (B) Electron microscopy images of V307R-BsFtsZ polymers in the absence and presence of 20 μm plumbagin under assembly conditions. Scale bars = 200 nm. (C) The amount of Pi released per mol V307R-BsFtsZ in the absence and presence of various concentrations of plumbagin.

Effects of plumbagin on assembly and GTPase activity of D199A-BsFtsZ

The polymerization profile of D199A-BsFtsZ was similar to that of wild-type FtsZ, except that the extent of light scattering was found to be higher than that of the control. Consistent with the results of the light-scattering experiment, D199A-BsFtsZ was found to form thicker polymers than those of WT-BsFtsZ (Figs 3B and 7B). The mean thicknesses of the polymers of WT-BsFtsZ and D199A-BsFtsZ were 18.4 ± 7.9 and 47.3 ± 23 nm, respectively. Docking analysis indicated that mutation of aspartic acid 199 to alanine may perturb the hydrogen bond between plumbagin and BsFtsZ. Furthermore, plumbagin (20 and 30 μm) did not inhibit assembly of D199A-BsFtsZ (Fig. 7A). In addition, electron microscopic analysis also suggested that plumbagin had no effect on assembly of D199A-BsFtsZ (Fig. 7B). Plumbagin did not alter the GTPase activity of D199A-BsFtsZ (Fig. 7C). Together, the results suggest that the variant D199A-BsFtsZ is resistant to the inhibitory effects of plumbagin.

Figure 7.

Effects of plumbagin on the assembly and GTPase activity of D199A-BsFtsZ. (A) The effect of plumbagin on the assembly kinetics of D199A-BsFtsZ (8 μm) was monitored in the absence (closed square) and presence of 2 μm plumbagin (closed circle), 5 μm plumbagin (closed upright triangle), 10 μm plumbagin (closed downward triangle), 20 μm plumbagin (closed leftward-pointing triangle) and 30 μm plumbagin (closed rightward-pointing triangle). (B) Transmission electron microscopy images showing the polymer morphology of D199A-BsFtsZ in the absence and presence of 20 μm plumbagin. Scale bar = 200 nm. (C) The effect of plumbagin on the GTPase activity of D199A-BsFtsZ (6 μm) was determined by measuring Pi release after 10 min of assembly.

Binding of plumbagin to WT-BsFtsZ, V307R-BsFtsZ and D199A-BsFtsZ

A hydrophobic fluorescent probe, 8-anilino-1-naphthalenesulfonic acid (ANS), has been widely used to determine the dissociation constant of ligand and protein interaction [24, 31, 32]. WT-BsFtsZ, V307R-BsFtsZ and D199A-BsFtsZ do not contain a tryptophan residue; therefore, the affinity of interaction of plumbagin with these proteins was determined using ANS. Analysis of the fluorescence change data yielded Kd values of 10.6 ± 0.9, 66.1 ± 18.7 and 95.9 ± 20 μm for WT-BsFtsZ, V307R-BsftsZ and D199A-BsFtsZ, respectively (Fig. S7). The results indicate that plumbagin interacts with V307R-BsftsZ and D199A-BsFtsZ much more weakly than with WT-BsFtsZ.

Discussion

In this study, plumbagin was found to bind to BsFtsZ and to reduce the assembly and GTPase activity of BsFtsZ in vitro. Interestingly, plumbagin did not discernibly inhibit the assembly and GTPase activity of D199A-BsFtsZ and V307R-BsFtsZ, indicating that D199A and V307R are likely to be important for its interaction with BsFtsZ. The affinity of interaction of plumbagin with V307R-BsftsZ and D199A-BsFtsZ was estimated to be much weaker than that of WT-BsFtsZ. Furthermore, plumbagin increased the length of B. subtilis 168 cells and perturbed the formation of Z-rings in bacteria without affecting nucleoid segregation, indicating that it inhibited bacterial cell division by perturbing the assembly of FtsZ.

Although plumbagin inhibited the assembly and GTPase activity of BsFtsZ, it had no discernable effect on either the assembly or GTPase activity of EcFtsZ, suggesting that the ligand-binding abilities of these proteins are different. Consistent with previous work [22], plumbagin did not inhibit the proliferation of E. coli cells. Resistance of E. coli cells to plumbagin treatment was previously proposed to be regulated by YgfZ and SodA [33, 34]; these proteins were suggested to degrade plumbagin into less toxic by-products and to remove the oxidative stress caused by plumbagin in E. coli [34]. In the present study, plumbagin was found not to inhibit the assembly or GTPase activity of EcFtsZ in vitro, which may be one of the reasons for its inactivity against E. coli cells (Fig. S5). This observation also suggests that it may be possible to selectively target a bacterial species by specifically inhibiting its FtsZ assembly.

Plumbagin did not inhibit the binding of TNP-GTP to FtsZ, but did inhibit the assembly of BsFtsZ in the presence of 1 mm GTP, indicating that it did not bind to the GTP-binding site on FtsZ. Moreover, in silico analysis indicated that the plumbagin-binding site on BsFtsZ is distinct from the GTP-binding domain of BsFtsZ. The plumbagin-binding site was found to be located towards the C-terminal domain of the FtsZ monomer, and also encompassed a significant region of the H7 helix (R191, Q192, Q195, G196 and D199). In addition, a few residues (e.g. N263 and T265) are located at the region immediately above the C-terminal interface of FtsZ, and are in contact with the nucleotide-binding domain of the next FtsZ monomer. This region of FtsZ has been suggested to be important for FtsZ polymerization [35]. Binding of plumbagin may alter the conformation of this region. Interestingly, some of the interacting residues of plumbagin and FtsZ were found to be similar to those of the recently discovered anti-staphylococcal compound PC190723 [26]. PC190723, a derivative of 3-methoxybenzamide (3-MBA), comprises two structural moieties, a benzamide group and a thiazolopyridine group. While the benzamide group was predicted to bind adjacent to residues R191, Q192, N263, V307 and T309, the thiazolopyridine group interacted with residues I172, E185, N188, I228 and I230. However, the two compounds had different effects on FtsZ assembly; plumbagin inhibited FtsZ assembly while PC190723 promoted assembly of FtsZ [36].

V307R-BsFtsZ displayed similar assembly kinetics and GTPase activity to WT-BsFtsZ, indicating that replacement of valine residue by an arginine at position 307 of BsFtsZ has no influence on the assembly pattern. However, the assembly characteristics of D199A-BsFtsZ were found to be different from those of WT-BsFtsZ. D199A-BsFtsZ polymerized to a greater extent and formed thicker bundles than WT-BsFtsZ. Recently, we found that an E93R mutation in EcFtsZ increased the assembly and bundling of EcFtsZ [37]. Computational studies indicated that the H7 helix participates in long-range interactions, implying a significant role in the folding and stability of the MjFtsZ (Methanococcus jannaschii) structure [35]. It also appears to be important in making contacts with different regions (e.g. the H1 helix and H1–S2 loop) in TmFtsZ (Thermotoga maritima) [35]. D199 is part of the H7 helix. The replacement of D199 by alanine may have affected the inter-domain contacts, resulting in altered morphology of polymers of the variant D199A-BsFtsZ.

Plumbagin did not inhibit the assembly and GTPase activity of V307R-BsFtsZ, indicating that V307 is important for the interaction of plumbagin with BsFtsZ. Replacement of the valine residue by an arginine residue may lead to partial loss of hydrophobic interaction with plumbagin; moreover, the arginine may also possibly occlude the plumbagin-binding site, resulting in the prevention of plumbagin binding by steric hindrance. Plumbagin also did not inhibit the assembly of D199A-BsFtsZ (Fig. 7A,B). The loss of the predicted hydrogen bond between D199 and plumbagin due to its substitution by alanine may be responsible for the resistance of this variant. Alternatively, the altered polymer morphology of this variant may contribute to the loss of sensitivity of D199A towards plumbagin. Therefore, V307 and D199 are likely to be important for the interaction of plumbagin and BsFtsZ. Expressing the variant proteins V307R-BsFtsZ and D199A-BsFtsZ in cells and studying the action of plumbagin on them in vivo will provide a better understanding of the involvement of the residues V307 and D199 in plumbagin binding, and this may be exploited in future studies for development of FtsZ-targeted potent antibacterial agents.

Experimental procedures

Plumbagin, PIPES, BSA, lysozyme, GTP, 4′,6-diamidino-2-phenylindole, phenylmethanesulfonyl fluoride, ANS, TNP-GTP and Cy3-conjugated goat anti-rabbit IgG were purchased from Sigma (St Louis, MO). Isopropyl thio-β-d-galactoside was obtained from Calbiochem (Darmstadt, Germany) Ni-NTA resin was obtained from Qiagen (Hilden, Germany), and Bio-Gel P-6 resin was purchased from Bio-Rad (Hercules, CA, USA). The rabbit polyclonal FtsZ antibody was produced by Bangalore Genei (Bangalore, India). All other reagents used for the study were of analytical grade.

Preparation of plumbagin solution

Plumbagin was soluble in 100% dimethylsulfoxide. The concentration of the stock solution of plumbagin was determined by measuring the absorbance at 418 nm in a V-530 UV/Vis spectrophotometer (JASCO, Tokyo, Japan). For each experimental set, the concentration of dimethylsulfoxide was kept at < 0.1%.

Effects of plumbagin on the growth of B. subtilis 168 cells

Bacillus subtilis 168 cells were grown in LB medium in the absence and presence of various concentrations (2, 5, 10 and 15 μm) of plumbagin at 37 °C. At 30 min intervals, the attenuance of the bacterial culture was measured at 600 nm in a JASCO V-530 UV/Vis spectrophotometer. The growth of B. subtilis 168 cells was monitored for 6 h. The experiment was repeated three times.

Determination of the minimum inhibitory concentration of plumbagin

The minimum inhibitory concentration of plumbagin for B. subtilis 168 and M. smegmatis cells was determined by the agar dilution method [28, 30]. Bacterial cells (1 × 105) were plated onto Luria agar plates containing various concentrations (5–50 μm) of plumbagin, and incubated overnight at 37 °C for B. subtilis 168 cells or 4 days for M. smegmatis cells. The experiment was performed three times. The concentration of plumbagin at which no colonies was observed on the plates was considered to be the minimum inhibitory concentration of plumbagin.

Visualization of bacterial morphology

Bacillus subtilis 168 cells were grown in the absence and presence of 5 μm plumbagin at 37 °C for 4 h. The cells were fixed using fixing solution containing 2.5% formaldehyde and 0.04% glutaraldehyde for 30 min. The fixed cells were pelleted at 18 000 g for 5 min at 4 °C, resuspended in NaCl/Pi, and diluted in 1:10 ratio. An aliquot (10 μL) of the cell suspension was then placed on a cover slip, and the morphology of the cells was visualized under a microscope (Eclipse TE2000-U microscope; Nikon, Tokyo, Japan) in differential interference contrast mode using a 60× objective. The images were captured using a CoolSNAP-Pro camera (Tucson, AZ, USA), and measurement of the bacterial cell length was performed using image pro plus software (Media Cybernetics, Silver Spring, MD, USA). For each experimental condition, ~ 300 cells were measured, and the mean cell length was calculated.

Immunofluorescence microscopy

Immunostaining was performed to monitor the effects of plumbagin on the Z-ring and nucleoids of B. subtilis 168 cells [24]. Briefly, B. subtilis 168 cells were grown in the absence and presence of plumbagin (5 and 15 μm) at 37 °C for 4 h, and subsequently fixed using 2.5% formaldehyde and 0.04% glutaraldehyde in NaCl/Pi for 30 min. The cells were permeabilized using 0.1% Triton X-100 in NaCl/Pi, and non-specific binding sites were blocked by incubating the cells at 37 °C in 2% BSA/NaCl/Pi. FtsZ was stained using rabbit polyclonal FtsZ antibody (1 : 50), and further stained using Cy3-conjugated goat anti-rabbit IgG (1 : 200). The nucleoids were stained using 20 μg·mL−1 4′, 6-diamidino-2-phenylindole (DAPI). The cells were visualized under a fluorescence microscope (Eclipse TE2000-U microscope) using a 60× objective. Analysis of the images was performed using image pro plus software. The frequency of Z-rings and nucleoids per micrometer of the cell length were determined on the basis of the results for ~ 300 cells.

Isolation and purification of wild-type BsFtsZ, variants of BsFtsZ and EcFtsZ

Recombinant WT-BsFtsZ was over-expressed in E. coli BL21(DE3) pLysS, and purified using a Ni-NTA column as described previously [28]. Briefly, cells containing the desired construct were grown at 37 °C in LB medium containing 100 μg·mL−1 ampicillin and 12.5 μg·mL−1 chloramphenicol, and induced with 1 mm isopropyl thio-β-d-galactoside for 6 h. The cells were harvested at 7780 g for 20 min at 4 °C, resuspended in lysis buffer A (50 mm NaH2PO4, 300 mm NaCl, pH 8.0 containing 2 mm PMSF and 1 mg.mL−1 lysozyme), and allowed to bind to a Ni-NTA column in the presence of 5 mm imidazole. The column was washed using 200 mL of buffer B (25 mm PIPES, 300 mm NaCl, pH 6.8) containing 50 mm imidazole at 4 °C, and then the protein was eluted using 50 mL of buffer B containing 250 mm imidazole at 4 °C, and subsequently desalted using Bio-Gel P-6 resin and stored at −80 °C.

The protein concentration was measured using Bradford reagent with BSA as the standard [38], and the final concentration was adjusted by including the correction factor of 1.2 for the FtsZ/BSA ratio [39]. The protein was subjected to high-speed centrifugation at 4 °C prior to each experiment to ensure removal of aggregates.

PCR-based site-directed mutagenesis of BsFtsZ was performed as described in the protocol supplied with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). Recombinant BsFtsZ-pET16b, B. subtilis ftsZ gene cloned in pET16b (Novagen, Darmstadt, Germany), was used as the template plasmid, and primers were designed with the purpose of individually replacing the aspartic acid residue at position 199 and the valine residue at position 307 by alanine and arginine residues, respectively. The sequences of primers used for the two mutations were 5′-GGGGTTCAAGGTATTTCTGCCTTGATTGCTACACCTGG-3′ and 5′-GCAATCAAGGCAGAAATACCTTGAACCCC -3′ for D199A, and 5′-GAAAATCTAAAAGATGAGATTCGGGTGACAGTGATTGCAACCG-3′ and 5′-CGGTTGCAATCACTGTCACCCGAATCTCATCTTTTAGATTTTC-3′ for V307R. The mutations were confirmed by DNA sequencing (performed by Xcelris Labs, Ahmedabad, India). The variant proteins were expressed in E. coli (BL21(DE3)) pLysS cells, and purified as for WT-BsFtsZ.

Recombinant EcFtsZ was over-expressed and purified from E. coli BL21 cells as described previously [40].

Light-scattering assay

WT-BsFtsZ, variants of BsFtsZ (V307R and D199A) and EcFtsZ in buffer C (25 mm PIPES, 50 mm KCl, pH 6.8) were separately incubated without or with various concentrations of plumbagin on ice for 15 min. Then 10 mm MgCl2 and 1 mm GTP were added to each of the reaction mixtures, and the cuvette was immediately transferred to an FP-6500 JASCO spectrofluorometer connected to a temperature-controlled water bath. The assembly kinetics were monitored by 90 ° light scattering [41, 42]. The light-scattering intensities of appropriate blanks were also measured and subtracted from each of the corresponding reaction sets to obtain corrected data. To verify whether Y273W-BsFtsZ was functionally active, Y273W-BsFtsZ (10 μm) was polymerized under assembly conditions (buffer C, 10 mm MgCl2 and 1 mm GTP) in the absence of plumbagin, and the polymerization trace was monitored by a 90 ° light-scattering assay at 500 nm.

Transmission electron microscopy

The effect of plumbagin on the polymer morphology of WT-BsFtsZ, variant BsFtsZs (D199A and V307R) and EcFtsZ was observed by transmission electron microscopy [25, 41]. BsFtsZ (6 μm) in buffer C was incubated in the absence and presence of 20 μm plumbagin on ice for 15 min. Following incubation, 10 mm MgCl2 and 1 mm GTP were added to each reaction mix and polymerized at 37 °C for 10 min. The polymeric suspension was then adsorbed onto formvar carbon-coated copper grids (300 mesh) and air-dried. The grids were then negatively stained with 2% uranyl acetate solution, dried extensively and observed under a Philips FEI Tecnai-G2 12 electron microscope (Amsterdam, Netherlands). A similar procedure was followed for sample preparation of variant BsFtsZs (D199A and V307R) and EcFtsZ.

Measurement of the GTPase activity of FtsZ

FtsZ (6 μm) in buffer C was incubated without and with various concentrations (3, 6, 12, 24 and 36 μm) of plumbagin on ice for 15 min. Then 10 mm MgCl2 and 1 mm GTP were added to the reaction mixtures, and the mixtures were immediately transferred to a 37 °C water bath. The hydrolysis reaction was quenched after 10 min by addition of 7 M perchloric acid (10% v/v). The amount of inorganic phosphate released was determined by the standard malachite green ammonium molybdate assay [25, 43].

Size-exclusion chromatography

The interaction between plumbagin and BsFtsZ was monitored using a G-25 column (BioRad, Hercules, CA, USA) of volume 25 mL [24]. The void volume of the column was found to be ~ 8.5 mL using blue dextran (2 mg·mL−1). When loaded separately on a pre-equilibrated column, BsFtsZ (15 μm) was found to elute within the void volume (~ 8.5 mL), and plumbagin (60 μm) eluted in the column volume (~ 34 mL). The flow rate of elution was maintained at 0.5 mL·min−1. The elution profile of the protein was monitored by measuring the absorbance at 595 nm using Bradford reagent [38], while the elution profile of plumbagin was monitored at 418 nm (λmax of plumbagin). BsFtsZ (15 μm) was incubated with plumbagin (60 μm) on ice for 30 min, and then the reaction mixture was loaded onto the pre-equilibrated G-25 column. The fractions were eluted again using 25 mm PIPES, pH 6.8. The absorbance of each aliquot was measured at 595 and 418 nm.

Determination of the dissociation constant for the interaction between plumbagin and Y273W-BsFtsZ

BsFtsZ lacks a tryptophan residue in its sequence; therefore we designed a variant Y273W-BsFtsZ containing a single tryptophan residue to monitor the binding interaction of plumbagin with BsFtsZ [28]. Y273W-BsFtsZ exhibited emission maxima at 333 nm. Furthermore, Y273W-BsFtsZ displayed similar assembly kinetics to that of native BsFtsZ, indicating that it is an active protein (Fig. S3).

Y273W-BsFtsZ (1 μm) was incubated in 25 mm PIPES (pH 6.8) without and with various concentrations (1, 2, 5, 10, 20, 30, 40 and 50 μm) of plumbagin at 25 °C for 15 min. The fluorescence spectra (310–350 nm) were collected using a cuvette of 0.3 cm path length by exciting the samples at 295 nm. Plumbagin showed no significant absorbance at either the excitation (295 nm) or the emission (333 nm) wavelength; therefore, no inner filter effect correction was required. The fluorescence spectra of the corresponding blanks were also recorded and subtracted from the individual sets to obtain corrected spectra. Plumbagin quenched the tryptophan fluorescence of Y273W-BsFtsZ in a concentration-dependent manner. The dissociation constant of the interaction between plumbagin and Y273W-BsFtsZ was determined using the equation:

display math

where ∆F is the change in the intrinsic tryptophan fluorescence intensity of Y273W-BsFtsZ when it interacts with plumbagin, ΔFmax is the maximum change in the tryptophan fluorescence intensity of Y273W-BsFtsZ when it is fully saturated with plumbagin, and L is the plumbagin concentration. The dissociation constant (Kd) of the interaction between plumbagin and Y273W-BsFtsZ was determined by fitting the fluorescence data using graphpad prism 5 software (Graph Pad Software, La Jolla, CA, USA).

Determination of the dissociation constant for the interaction between WT-BsFtsZ, V307R-BsFtsZ, D199A-BsFtsZ and plumbagin using ANS

WT-BsFtsZ and the variants V307R-BsFtsZ and D199A-BsFtsZ (2 μm) were incubated with varying concentrations of plumbagin at 25 °C for 10 min. Then the reaction mixtures were incubated with 30 μm ANS for 20 min at 25 °C. The fluorescence spectra were collected in the range of 430–520 nm by exciting the samples at 350 nm. A cuvette of 0.3 cm path length was used to reduce the inner filter effect. Plumbagin was found to decrease the fluorescence of the FtsZ–ANS complexes in a concentration-dependent manner. As plumbagin shows absorbance at 418 nm, the fluorescence spectra were corrected for the inner filter effect [44] according to the equation:

display math

where Fcorr is the corrected fluorescence, Fobs is the observed fluorescence, and λex and λem are the absorbance of plumbagin at the excitation and emission wavelengths of ANS, respectively. The Kd for the interaction of plumbagin with FtsZ was determined by fitting the fluorescence data to a binding equation as described above.

Circular dichroism spectroscopy

BsFtsZ (5 μm) in 10 mm phosphate buffer (pH 6.8) was incubated without and with various concentrations (5, 10 and 20 μm) of plumbagin at 25 °C for 15 min. The far-UV CD spectra were monitored using a cuvette of 0.1 cm path length in the wavelength range 200–260 nm using a JASCO J810 spectropolarimeter. CD spectra of appropriate blanks were taken and subtracted from the corresponding datasets to obtain corrected spectra [37].

Effect of plumbagin on the binding of TNP-GTP to BsFtsZ

TNP-GTP, a fluorescent analog of GTP, has been reported to bind to the GTP site on FtsZ [25, 28, 45]. BsFtsZ (6 μm) in 25 mm PIPES, pH 6.8, containing 10 mm MgCl2, 200 mm NaCl and 10% glycerol was incubated with and without 20 or 40 μm plumbagin on ice for 15 min. Then the reaction mixtures were incubated with 15 μm TNP-GTP for 2 h. The emission spectra (500–600 nm) were recorded using 410 nm as the excitation wavelength in a cuvette of 0.3 cm path length. Corresponding blank spectra for samples containing only plumbagin and TNP-GTP were also recorded. The fluorescence data were corrected for the inner filter effect [44]. The effect of GTP on the binding of TNP-GTP to FtsZ was similarly monitored.

Docking methodology

Autodock 4 was used for docking [46]. The structure of BsFtsZ was obtained from the Protein Data Bank (PDB ID 2VXY) [26, 47], and the plumbagin structure was generated using the PRODRG server [48]. The entire FtsZ molecule was enclosed in a grid box of 170 × 170 × 170 grid points with a grid spacing of 0.375 Å. The algorithm employed was Lamarckian genetic algorithm with default parameters. Twenty docking jobs were performed, each of 100 runs, such that 2000 output conformations were obtained. All the conformations were clustered using a cut-off root mean square deviation of 1 Å. The cluster with the maximum number of conformations and the lowest binding energy was chosen as the binding site of plumbagin on BsFtsZ.

Similar methodology was used for docking of EcFtsZ and plumbagin. As the crystal structure of EcFtsZ is not available in the Protein Data Bank, we modeled the structure of EcFtsZ using the CHPmodels protein homology modeling server [49]. The modeled structure was checked for geometry by PROCHECK analysis [50] and subsequently used for docking of plumbagin.

Acknowledgements

This work was supported by a grant from the Indian Council of Scientific and Industrial Research (to D.P.). We thank the Centre for Research in Nanotechnology and Science, Indian Institute of Technology Bombay, for the electron microscopy facility, and Ankit Rai and Sonia Kapoor for critically reading the manuscript.

Ancillary