Vanadate inhibits Feo-mediated iron transport in Vibrio cholerae

Abstract

Iron is an essential element for Vibrio cholerae to survive, and Feo, the major bacterial system for ferrous iron transport, is important for growth of this pathogen in low-oxygen environments. To gain insight into its biochemical mechanism, we evaluated the effects of widely used ATPase inhibitors on the ATP hydrolysis activity of the N-terminal domain of V. cholerae FeoB. Our results showed that sodium orthovanadate and sodium azide effectively inhibit the catalytic activity of the N-terminal domain of V. cholerae FeoB. Further, sodium orthovanadate was the more effective inhibitor against V. cholerae ferrous iron transport in vivo. These results contribute to a more comprehensive biochemical understanding of Feo function, and shed light on designing effective inhibitors against bacterial FeoB proteins.

Graphical Abstract

Graphical Abstract.

Sodium orthovanadate inhibits ferrous iron transport in the human pathogen Vibrio cholerae, likely via inhibition of the ATPase activity of FeoB.

Introduction

Iron is an essential element for almost all microorganisms to survive. Its transport and metabolism are crucial in bacteria.¹ Feo is the major bacterial ferrous iron transport system, and it has been shown to be associated with virulence in several bacterial pathogens.^2–5 FeoB, an essential component of the Feo system, is an iron permease possessing nucleotide hydrolysis activity (mostly GTPase) in its amino-terminal cytosolic domain.⁶ The hydrolysis activity is required for iron transport, and it may function by regulating iron uptake or directly providing the energy for iron transport across the membrane. Deletion or mutation of key residues in its nucleotide hydrolysis domain causes a complete loss of its iron transport activity, indicating that NFeoB may be a good target for novel antibacterial therapeutics.^6,7

The Feo system was first described in Escherichia coli, and FeoB was initially predicted to hydrolyze ATP based on its sequence homology with nucleotide-binding sites of ATPases.⁸ This was supported by evidence in Helicobacter pylori showing that the addition of ATP hydrolysis inhibitors resulted in abrogation of ferrous iron transport in that species.³ However, studies of E. coli FeoB demonstrated that the N-terminus contained characteristics of G proteins and that it hydrolyzes GTP but does not bind or hydrolyze ATP.^7,9 Therefore, FeoB has been reclassified as a GTPase.

Vibrio cholerae is a human pathogen that causes a severe diarrheal disease associated with devastating epidemics. It has an absolute requirement for iron, and acquisition of ferrous iron is likely to play a role in anoxic niches in the lumen of the human intestine.¹⁰ In a previous study, we reported that V. cholerae FeoB possesses both ATP and GTP hydrolysis activity, although the protein has a GTPase domain that closely resembles eukaryotic G proteins.⁶ However, biochemical details of FeoB NTPase activity and its role in iron transport via FeoB are still unclear. In this study, we evaluated widely used ATPase inhibitors on the ATPase activity of the N-terminal, cytosolic domain of V. cholerae FeoB (VcNFeoB) and hypothesized the possible inhibition of NFeoB via ATPase inhibitors.

Materials and methods

Bacterial strains and growth conditions

The bacterial strains and plasmids used in this study were reported in Shin et al.⁶Escherichia coli and V. cholerae strains were grown in Luria–Bertani (LB) medium at 37°C. The following antibiotics were used at the indicated concentrations: ampicillin, 25 µg/mL, and kanamycin, 50 µg/mL. Sodium orthovanadate, sodium azide, and carbonyl cyanide m-chlorophenyl hydrazine (CCCP) were purchased from Sigma-Aldrich (St. Louis, MI, USA).

Molecular docking simulation

For the molecular docking prediction, the E. coli cytosolic domain of FeoB (EcNFeoB, Protein Data Bank (PDB) ID: 3I92) and homology models of VcNFeoB (built from the N-terminal portion of Klebsiella pneumonia NFeoB [KpNFeoB] as a template; PDB ID: 2WIC) and Staphylococcus aureus N-terminal FeoB (SaNFeoB; built from Streptococcus thermophilus as a template; PDB ID: 3SS8) were used. The target proteins were prepared for docking based on the dock prep tool in UCSF Chimera.¹¹ As ligand molecules, structures of 31 commonly used ATPase inhibitors were tested (Table S1).

The docking simulation was performed by the AutoDock Vina program, which is an open source tool for molecular docking and virtual screening.¹² A grid box for AutoDock Vina (30 × 30 × 30 Å for VcNFeoB and SaNFeoB, and 40 × 40 × 40 Å for EcNFeoB) was generated by Autodock Tools, and the grid was centered at x-, y-, and z-axes coordinates of −0 × 20 × −30 for VcNFeoB and SaNFeoB and 60 × −15 × 80 for EcNFeoB, respectively. The simulation results were visualized using the UCSF Chimera ViewDock extension, and the energy scores were compared for each ligand based on the Vina algorithm.¹²

Protein expression and purification

A pET21a-based expression vector encoding the cytosolic N-terminus of FeoB from V. cholerae (amino acids 1–272, VcNFeoB) with a hexahistidine tag was generated. VcNFeoB was expressed in E. coli BL21 from this construct and purified following a protocol published previously.⁶

Nucleotide hydrolysis assay

A malachite-green-based colorimetric assay was used to measure inorganic phosphate released through VcNFeoB's nucleotide hydrolysis activity.^6,13,14 For determination of kinetic parameters, the initial rate data were fitted to the Michaelis–Menten equation by nonlinear regression, and K_I was calculated from a global fit of the data to an inhibition model using GraphPad Prism 9 (San Diego, CA, USA).

Radioactive iron transport assay

Vibrio cholerae strain EPV6 harboring the indicated plasmid was grown in LB broth for 2.5 h. Cells were pelleted, washed twice with M9 salts, and incubated in M9 minimal medium with 50 μg/ml ethylenediamine-N,N′-diacetic acid (EDDA; wt/vol) for 2 h. Cells were pelleted, washed twice with M9 salts, and resuspended in M9 minimal medium with 5 mM sodium ascorbate and/or 1 mM inhibitors at a concentration of 1.5 (optical density at 650 nm) per milliliter of medium. After adding inhibitors, cells were incubated for 15 min at room temperature. The assay was conducted by adding radioactive ⁵⁵FeCl₂ (Perkin-Elmer Life and Analytical Sciences) to the iron-deprived bacterial cell cultures.¹⁵ Cells were collected by filtration, and the amount of radioactive iron was determined by scintillation counting.⁶

Results and discussion

In silico molecular docking study predicts that VcNFeoB would interact with P-type ATPase inhibitors

The scoring function of molecular docking prediction comes with a prediction of a geometry, which allows for inhibitor screening in the context of a binding site.¹⁶ By the function, we can easily identify new active compounds toward a particular target protein with reduced time and expenses.¹⁷ In fact, a number of recent studies have combined the technologies of structure-based virtual screening and in vitro experimental validations for discovery of novel classes of drug candidates.^17,18

Previous studies have shown that several FeoB proteins, such as V. cholerae and S. aureus FeoB, possess ATP and GTP hydrolysis activity, while E. coli FeoB catalyzes only GTP hydrolysis.^6,13 ATP hydrolysis by FeoB differs from GTP hydrolysis on the mono- and divalent cation requirements. To further gain an insight into the characteristics of the ATP hydrolysis activity in FeoBs, we first performed an in silico docking simulation of common ATPase inhibitors that target specific types of ATPases (P-, F-, and V-types) against the structural homology model of VcNFeoB (Table S1).

Among 31 ATPase inhibitors, citreoviridin, omeprazole, and lansoprazole showed the highest binding scores (−8.2, −7.7, and −7.4 kcal/mol, respectively). Regarding the type of ATPase inhibitors, the P-type inhibitors had comparably lower binding scores than other types of inhibitors, indicating a probably strong binding interaction of inhibitors in this class with VcNFeoB (Fig. 1A). The P-type ATPase inhibitors (lansoprazole and omeprazole) occupied the active catalytic site of ATP hydrolysis in VcNFeoB (Fig. 1B,C). Omeprazole and lansoprazole are antisecretory compounds that prevent gastric acid secretion by selective inhibiting the H⁺/K⁺-ATPase system.¹⁹ These molecules bind covalently to cysteine residues on parietal H⁺/K⁺-ATPase, resulting in stable disulfide bonds.²⁰ Both inhibitors have similar chemical structures as polycyclic aromatic compounds containing a sulfinyl group attached at the position 2 of a benzimidazole moiety. Citreoviridin is a toxic secondary metabolite produced by Penicillium and Aspergillus species, which is known to inhibit the activity of F-type ATPases. The inhibitor interacts with a region near the β-subunit arginine 398 of F₁ATPase that lies at the α/β-interface.²¹ The citreoviridin binding causes a reduced rotation rate and increases the duration of the catalytic dwell for the ATPase activity.

Fig. 1.

In silico docking prediction of ATPase inhibitors against VcNFeoB. (A) Predicted binding scores depending on the types of inhibitors indicate that P-type ATPase inhibitors would tightly interact with the enzyme. (B–D) Molecular docking of GTP analog (B), lansoprazole (C), and omeprazole (D) to the GTP binding site of VcNFeoB. The inhibitors are shown as balls and sticks. Asterisks represent statistical significance below P < 0.05 (*) by one-way ANOVA and Tukey's multiple comparison test. The error bars represent the standard deviations.

To extend these findings to the FeoBs present in other types of bacteria, we predicted the binding scores of ATPase inhibitors on the N-terminal domain of S. aureus (SaNFeoB), which possesses even higher ATPase activity than VcNFeoB,¹³ and that of E. coli (EcNFeoB), which has GTPase activity only. SaNFeoB, but not EcNFeoB, showed a similar pattern to VcNFeoB. The higher preference of ATPase activity in SaNFeoB may suggest that this protein would be prone to more tightly interact with the P-type ATPase inhibitors than EcNFeoB, which is solely GTPase (Fig. 2). Overall, the P-type ATPase inhibitors showed better interaction with the docked residues in VcNFeoB, although the molecular docking prediction may have limitations on the lack of confidence in the ability of scoring functions to give accurate molecular interaction that sometimes does not correlate well with the experimental binding affinities.²²

Fig. 2.

In silico docking prediction of ATPase inhibitors against SaNFeoB (A) and EcNFeoB (B). Predicted binding scores depending on the types of inhibitors indicated that P-type ATPase inhibitors would tightly interact with SaNFeoB but not with EcNFeoB. Asterisks represent statistical significance below P < 0.05 (*) and 0.01 (**) by one-way ANOVA and Tukey's multiple comparison test. The error bars represent the standard deviations.

Sodium orthovanadate and sodium azide inhibit ATP hydrolysis of VcNFeoB

To further identify the mechanism of ATP hydrolysis, we determined the effects of ATPase inhibitors, sodium orthovanadate (OV, P-type inhibitor), sodium azide (NaN₃, F-type inhibitor), dicyclohexylcarbodiimide (DCCD, F-type inhibitor), and N-ethylmaleimide (NEM, V-type inhibitor), on the ATPase activity of VcNFeoB. Using 50% inhibition as the cutoff, the ATPase inhibitors OV and NaN₃ effectively inhibited ATPase activity (Fig. 3A). Unlike ATP hydrolysis, OV and NaN₃ decreased the GTP hydrolysis activity only by 38% and 30%, respectively, suggesting that the catalytic site in VcNFeoB for ATP and GTP hydrolyses may be different (Fig. 3B).

Fig. 3.

Four inhibitors widely used for blocking ATP hydrolysis were tested for inhibition of ATP (A) and GTP (B) hydrolysis by VcNFeoB. ATPase and GTPase activity are expressed as percentage activity relative to the basal activity registered with H₂O (OV and NaN₃) or ethanol (DCCD and NEM). Asterisks represent statistical significance P < 0.0001 (****) by one-way ANOVA and Šidák's multiple comparison test with more than 50% inhibition. The error bars represent the standard deviations for triplicates. OV, sodium orthovanadate; NaN3, sodium azide; DCCD, N,N-dicyclohexylcarbodiimide; and NEM, N-ethylmaleimide.

Based on our findings with OV and NaN₃, we conducted enzyme kinetic analyses to determine the potential inhibition mechanism by which these compounds affect VcNFeoB (Fig. 4). It is noted that although the docking prediction of citreoviridin gave a high score as one of the F-type ATPase inhibitors, we evaluated sodium azide instead for the in vitro inhibition assay. In the current study, we hypothesized that the phosphate-mimicking molecules, which are OV and azide, would be structurally related while slightly different in the coordination for inhibiting the phosphoryl group transfer. Also, because of the structural simplicity of the two molecules, the molecular docking prediction on OV and azide could not be evaluated.

Fig. 4.

Sodium orthovanadate (A) and sodium azide (B) inhibit the ATP hydrolysis activity of VcNFeoB. The reaction rates for ATP hydrolysis were measured as described in the section Materials and methods. The initial rate data were fitted to the Michaelis–Menten equation by nonlinear regression. The error bars represent the standard deviations.

Assuming the possible interaction of OV and azide with the catalytic site of NFeoB, we further tested the effects of the two molecules.

OV and NaN₃ inhibited ATPase activity of VcNFeoB in a dose-dependent manner with ranges of 100–2000 and 10–1000 μM, respectively. The nonlinear regression plots of velocity versus nucleotide concentration indicate that OV and NaN₃ are mixed model inhibitors with K_I = 538.7 and 151.5 μM for ATP, respectively (Figs. S1 and S2). It is noted that both of the inhibitors showed similar ranges of regression values for the goodness of fit among competitive, noncompetitive, and mixed inhibitions in accordance with the linear regression using the double reciprocal plots. In particular, mixed inhibition gave the best fitness values, although the differences among the inhibition modes were slight. Although the binding of vanadate is shown to occur at the phosphate sites competing with the substrate ATP, a number of studies have shown that the inhibition by vanadate was apparently noncompetitive or uncompetitive versus ATP.^23–27 The noncompetitive type of vanadate inhibition could be by the formation of an enzyme–ADP–vanadate complex analogous to the transition state during catalysis. Our results also exhibited the mixed and noncompetitive inhibitions as the better fitted models than others despite the slight differences among the inhibition modes. To provide the details of the inhibition mechanism of OV and NaN₃, elaborate pre-steady-state kinetic analysis is highly required.

Vanadate generally acts as a phosphate analogue, presumably by inhibiting ATPases as a transition state analogue for transfer of a phosphoryl group.²⁸ It is widely known as a P-type ATPase inhibitor and has a tetrahedral geometry, so it may act as an analogue of the transition state in phosphoryl transfer reactions.^29,30 More specifically, a recent study evaluated transition state structures of vanadium–phosphatase protein complexes and found that 5-coordinate vanadium in protein complexes was trigonal bipyramidal with umbrella distortions for the phosphoryl group transfer reactions.³¹ In contrast, NaN₃ inhibits F-type ATPases by interacting with the β-phosphate of the ADP molecule and with residues in the ADP-binding catalytic subunit, occupying a position similar to that of the γ-phosphate of ATP.³² Bowler et al. suggest that azide may inhibit several other types of P-loop ATPases too.³² The inhibition by these substrate analogues suggests that VcNFeoB shares a common structural conformation on its catalytic site with archetypical ATPases.

CCCP inhibits nucleotide hydrolysis of VcNFeoB

CCCP is a protonophore able to dissipate membrane potential and perturbate proton motive force in bacteria.³³ In general, the sodium–potassium pump activity of P-type ATPases is affected by protonophores.³⁴ Additionally, a study on H. pylori FeoB reported that ferrous iron transport was inhibited in this species by the protonophore carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) as well as ATPase inhibitors, suggesting an active iron transport mechanism energized by ATP.³ However, the enzymatic mechanism of FeoB inhibition by the protonophore has not been explained yet.

We investigated the effect of CCCP on the catalytic activity of VcNFeoB to verify if the protonophore is also able to inhibit nucleotide hydrolysis of VcNFeoB. As shown in Fig. 5A, 1 mM CCCP significantly inhibited ATP hydrolysis (2.1-fold reduction). Interestingly, CCCP was effective in inhibiting GTP hydrolysis as well, leading to a 4.6-fold reduction. To predict the molecular interaction between VcNFeoB and CCCP, we docked CCCP into the structural model of VcNFeoB, obtaining −7.0 kcal/mol as the best score. The model of the VcNFeoB–inhibitor complex revealed that CCCP could interact with the enzyme in the active site (Fig. 5B). CCCP is widely known as a proton motive force inhibitor, uncoupling the proton gradient that is established in the electron transport chain and reducing the ability of ATP synthase to function. However, it is also known to directly bind to several enzymes and inhibit their activity, which are not associated with ATPases such as E. coli transcriptional repressor EmrR³⁵ and Candida biodinii alcohol oxidase.³⁶ In the current study, we found that CCCP inhibited the cytosolic domain of VcFeoB by chance and assume that the inhibition would be independent of the mechanism of proton motive force. Further structural analysis is required to elucidate the interaction of residues in VcNFeoB with CCCP and their effects on the enzyme mechanism.

Fig. 5.

CCCP inhibits the ATP and GTP hydrolysis activities of VcNFeoB. (A) The reaction rates for NTP hydrolysis were measured as described in the section Materials and methods. Asterisks represent statistical significance below P < 0.01 (**) and 0.001 (***) by one-way ANOVA and Dunnett's multiple comparison test. The error bars represent the standard deviations. (B) Molecular docking of CCCP to the ATP binding site of VcNFeoB. The inhibitors are shown as balls and sticks.

Sodium orthovanadate inhibits V. cholerae ferrous iron transport system mediated by FeoB

To test the effect of the inhibitors on bacterial iron transport in vivo, we carried out functional iron transport assays. These studies were done in V. cholerae EPV6, which carries null mutations in all its inorganic iron transport systems.³⁷ Introducing a plasmid-encoded iron transport system, such as Feo, allows wild-type growth of the EPV6 strain. Thus, iron uptake via Feo can be studied in the absence of background from other iron transporters.^37,38 In addition, we have generated a plasmid encoding the feo operon with the D122N mutation in feoB, which disrupts the GTPase activity but not the ATPase activity,⁶ allowing us to determine the effects of inhibitors on the ATPase activity alone. As shown in Fig. 6, of the three molecules effective against in vitro nucleotide hydrolysis activity of VcNFeoB, only OV inhibited ferrous iron transport in vivo (32.3% compared with the vehicle control). Combination of OV and NaN₃ produced a similar result (34.9%). While the EPV6 strain carrying the empty vector was not affected by OV, the strain with the D122N substitution showed decreased iron transport activity in the presence of OV. This indicates that the ATP hydrolysis of the D122N mutant is sufficient for Feo iron transport and inhibiting this ATPase activity with OV prevents iron uptake.

Fig. 6.

Sodium orthovanadate inhibits ferrous iron transport in V. cholerae EPV6 expressing the functional wild-type Feo system (pFeo-WT) and a mutant Feo system with disrupted GTPase activity (pFeo-D122N), but not in the EPV6 control (pWKS30, the empty vector). Effect of CCCP, OV, and NaN₃ (1 mM for each inhibitor) on in vivo iron transport was tested, showing a significant inhibition of the transport by addition of OV or OV + NaN₃. In the absence of the plasmid encoding the Feo system (pWKS30), iron transport in EPV6 was less than 20% of the EPV6 + pFeo-WT and not affected by OV. Asterisks represent statistical significance below P < 0.05 (*) by comparison with the vehicle control. The error bars represent the standard deviations. CCCP, carbonyl cyanide m-chlorophenylhydrazone; OV, sodium orthovanadate; and NaN3, sodium azide.

Velayudhan et al. reported that the protonophore FCCP and ATPase inhibitors, such as vanadate and DCCD, effectively inhibit in vivo ferrous iron transport in H. pylori.³ Our results showed consistent inhibitory activity by OV between in vitro and in vivo settings. However, unlike FCCP, treatment with CCCP did not affect iron transport, and this might be caused by slight differences in the inhibitor structure or in membrane permeability for V. cholerae. Moreover, sodium azide did not alter iron transport by Feo despite its function as an ATPase inhibitor. The mechanism of in vivo accessibility and action of these molecules on the Feo system is still unclear, but we hypothesize that the roles of CCCP and sodium azide in dispersing the bacterial membrane potential, as a protonophore and a H⁺-translocating F-type ATPase inhibitor, respectively, would affect the Feo-mediated ferrous iron transport.^39,40 To investigate the specific mechanism of iron transport inhibited by OV, it will be necessary to define the effects of OV on global bacterial metabolism and energy generation.

With respect to the inhibitory effect of OV on bacterial iron uptake suggested as an antimicrobial candidate, the vanadium compounds are also widely associated in many biological systems, so they have been proposed for the treatment of metabolic diseases, including diabetes and cancer.^41–43 In particular, the molecules activate numerous signaling cascades and transcriptional factors due to their similarity with phosphate. Careful investigation of the OV effects on host metabolism would maximize the effectiveness of the molecule for prevention of infection as well as metabolic disorders.

Conclusion

In this study, we have shown that ATPase inhibitors, structural analogues for the conformation of the phosphate group, inhibited the adenine nucleotide hydrolysis activity of VcNFeoB. CCCP also directly reduced both the ATP and GTP hydrolysis activity of VcNFeoB. Of the three inhibitors, only OV inhibited ferrous iron transport in V. cholerae. Our results support the hypothesis that VcNFeoB ATP hydrolysis activity alone is sufficient for Feo iron transport function in V. cholerae and other species that have dual ATP/GTPase-type FeoB.^6,13 In addition, we found CCCP as a novel inhibitor against VcNFeoB, although it did not affect the iron transport activity in vivo. CCCP is an uncoupler of oxidative phosphorylation and binds to the cytochrome c oxidase complex.⁴⁴ The current result indicates that the molecule would also be involved in many other biological activities in prokaryotes besides oxidative phosphorylation. Combined with our previous findings on screening FeoB inhibitors,¹⁴ further research on designing an effective inhibitor against bacterial FeoBs is under investigation.

Author contributions

M.S. and S.M.P. wrote the first draft of the manuscript. M.S. and C.G. performed the laboratory analysis, analyzed the data, and prepared the figures and tables. M.S., C.G., and S.M.P. edited the manuscript. All the authors read and approved the final version.

Funding

This research was supported by the Basic Science Research Program (NRF-2019R1I1A1A01058125) through the National Research Foundation of Korea and by the Cooperative Research Program for Agriculture Science & Technology Development (PJ016227012021) through the Rural Development Administration, Republic of Korea, to M.S., and National Institutes of Health grant R01AI091957 to S.M.P.

Conflicts of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability

The data underlying this article are available in the article and in its online supplementary material.