The enterobactin biosynthetic intermediate 2,3‐dihydroxybenzoic acid is a competitive inhibitor of the Escherichia coli isochorismatase EntB

Abstract

The Escherichia coli enterobactin biosynthetic protein EntB is a bifunctional enzyme that catalyzes hydrolysis of isochorismate via its N‐terminal isochorismatase (IC) domain, and then transfers phosphopantetheinylated 2,3‐DHB to EntF via the EntB C‐terminal aryl carrier protein (ArCP) domain. Here we used a fluorescence anisotropy binding assay to investigate the ability of 2,3‐DHB to bind to enzymes in the DHB synthetic arm of the pathway. We found that 2,3‐DHB binds to EntE as a natural substrate with high affinity (K _D = 0.54 μM). Furthermore, apo‐EntB was found to bind to 2,3‐DHB with moderate affinity (K _D = 8.95 μM), despite the fact that this intermediate is neither a substrate nor a product of EntB. Molecular docking simulations predicted a top‐ranked ensemble in which 2,3‐DHB is bound at the isochorismatase active site of apo‐EntB. Steady‐state coupled enzymatic assays revealed that 2,3‐DHB is a competitive inhibitor of apo‐EntB isochorismatase activity (K _i ~ 200 μM), consistent with modeling predictions. Monitoring the EntC–EntB coupled reaction in real time via isothermal titration microcalorimetry confirmed that EntB was required to drive the EntC reaction toward isochorismate formation. Furthermore, addition of 2,3‐DHB to the ITC‐monitored reaction resulted in a suppression of integrated reaction heats, consistent with our observation that the molecule acts as a competitive inhibitor of EntB. Finally, we found that 2,3‐DHB lowered the efficiency of EntC–EntB isochorismate channeling by approximately 70%, consistent with steric blockage of the isochorismatase active site by bound 2,3‐DHB. Given its inhibitory properties, we hypothesize that 2,3‐DHB plays a regulatory role in feedback inhibition in order to maintain iron homeostasis upon intracellular accumulation of sufficient ferric enterobactin.

1. INTRODUCTION

Siderophores are iron‐chelating molecules secreted by numerous prokaryotic species as well as some eukaryotes in order to scavenge scarce extracellular ferric iron under aerobic conditions (Neilands, 1995; Renshaw et al., ²⁰⁰²). Enterobactin is a catecholate siderophore with extraordinarily high affinity for ferric iron (K _D ~ 10⁻³⁵ M at physiological pH) (Carrano & Raymond, 1979). It is comprised of three 2,3‐dihydroxybenzoic acid (2,3‐DHB) subunits connected via a triserine trilactone core. While 2,3‐DHB itself has iron‐chelating properties, the three‐dimensional triscatechol arrangement of three DHB moieties in enterobactin confers optimal affinity for Fe³⁺ (K _A = 10⁵² M⁻¹) (Harris et al., 1979; Raymond et al., ²⁰⁰³). In addition to being an enterobactin precursor, 2,3‐DHB has also been reported to be a precursor for other siderophores (vibriobactin and anguibactin), and to function as a secondary metabolite in plants as a response to pathogen infection (Bartsch et al., 2010; Marín et al., ²⁰¹²).

Since over‐accumulation of iron leads to cellular damage, biosynthesis of enterobactin is tightly regulated (Buss et al., 2001). Expression of the genes encoding enterobactin biosynthetic enzymes (entC, entB, entA, entE, entF, entD, and entH) is controlled by the protein Ferric Uptake Regulator (Fur). In iron‐replete conditions, holo‐Fur binds to the promoters of three ent gene clusters, silencing gene expression. Upon iron starvation, apo‐Fur dissociates from the promoters, resulting in up‐regulation of ent gene products and subsequent enterobactin biosynthesis (Andrews et al., 2003; Brickman et al., ¹⁹⁹⁰). In Escherichia coli, enterobactin is synthesized in the cytoplasm in two stages. In the first stage (Figure 1), the biosynthetic intermediate 2,3‐DHB is produced from chorismate by sequential reactions catalyzed by three enzymes: (i) EntC (isochorismate synthase; EC 5.4.4.2), (ii) the N‐terminal isochorismatase (IC) domain of EntB (EC 3.3.2.1), and (iii) EntA (2,3‐dihydro‐2,3‐dihydroxybenzoate dehydrogenase; EC 1.3.1.28). In the second stage of the pathway, 2,3‐DHB is adenylated by EntE (2,3‐dihydroxybenzoyl‐AMP ligase; EC 6.2.1.71) in an ATP‐dependent reaction in order to facilitate its covalent attachment to the aryl carrier protein (ArCP) domain of EntB and subsequent delivery to phosphopantetheinylated EntF (enterobactin synthase). Enterobactin production from activated 2,3‐DHB then involves three EntF‐catalyzed cycles of condensation with L‐serine via the formation of non‐ribosomal peptide bonds followed by cyclization and product release (Lai et al., 2006). Enterobactin is produced in the cytoplasm and secreted to the extracellular environment by a TolC‐dependent process (Corinna et al., 2005). Following secretion, enterobactin chelates extracellular ferric iron, and the ferric‐enterobactin complex is then imported via TonB‐dependent uptake (Raymond et al., 2003).

FIGURE 1.

Reaction scheme. (a) Reactions catalyzed by EntB, EntC, and EntA. (b) Reaction catalyzed by MenD. SEPHCHC: 2‐succinyl‐5‐enolpyruvyl‐6‐hydroxy‐3‐cyclohexene‐1‐carboxylate.

In order to deliver 2,3‐DHB to holo‐EntF, the ArCP domain of apo‐EntB itself must be phosphopantetheinylated at Ser 245 by the accessory enzyme EntD to form holo‐EntB. The catalytic efficiency of phosphopantetheinylation at the EntB C‐terminal domain is insensitive to the presence of the N‐terminal IC domain (Gehring et al., 1997). Similarly, the N‐terminal isochorismatase activity of EntB has been shown to be insensitive to the phosphopantetheinylation status of the C‐terminal ArCP domain (Jiang & Guo, 2007). In agreement with these observations, our group has reported that enterobactin biosynthesis and secretion can still occur in an E. coli entB ⁻ knockout strain transformed with two expression constructs under Fur control that encoded discrete functional IC and ArCP domains (Pakarian & Pawelek, 2016).

The bifunctional nature of EntB, as well as its promiscuity in forming pairwise interactions with other enzymes in the pathway (EntC, EntA, EntE, EntF, Drake et al., 2006; Khalil & Pawelek, ²⁰⁰⁹; Lai et al., ²⁰⁰⁶; Ouellette et al., ²⁰²²; Pakarian & Pawelek, ²⁰¹⁶) highlights the central role EntB plays in enterobactin biosynthesis. We consider EntB to be a protein–protein interaction (PPI) hub, playing not only a catalytic role but also a structural role in promoting the formation of a biosynthetic multienzyme complex similar to the pyoverdine siderosome reported in Pseudomonas aeruginosa (Imperi & Visca, 2013). In support of this, we have produced in vivo crosslinking results demonstrating that EntC, EntB, EntA, and EntE co‐migrate as a cross‐linked species isolated from iron‐depleted E. coli cell lysates (unpublished data). We have also recently reported evidence of leaky isochorismate channeling between EntC and EntB in vitro (Bin & Pawelek, 2024), demonstrating that pairwise PPIs between Ent enzymes can function to optimize metabolite flux in the direction of 2,3‐DHB formation.

Given the central role played in enterobactin biosynthesis, EntB is a likely target for regulation. In living organisms, accumulation of intracellular iron is toxic due to the Fenton reaction and production of reactive oxygen species leading to cell death (Dixon & Stockwell, 2014). Regulation of enterobactin biosynthesis in E. coli is essential for maintaining iron homeostasis, especially following attainment of sufficient intracellular iron. It has long been known, for example, that in an iron‐rich environment the genes necessary for enterobactin biosynthesis are repressed by holo‐Fur (Lucía et al., 1999). However, beyond Fur repression, our current understanding of how enterobactin biosynthesis is regulated is highly limited. Here we report novel evidence of competitive inhibition of EntB by 2,3‐DHB, which is neither a substrate nor a product of this enzyme. Given such an inhibitory role, we hypothesize that 2,3‐DHB can additionally serve as a negative feedback regulator of iron homeostasis.

2. MATERIALS AND METHODS

2.1. Reagents

2,3‐DHB, porcine lactate dehydrogenase (LDH) and chorismic acid were purchased from Sigma‐Aldrich (St. Louis, MO). All other reagents were purchased from Bioshop Canada, Inc. (Burlington, ON).

2.2. Molecular docking and analysis

Atomic coordinates of 2,3‐DHB (ligand) and the dimeric EntB x‐ray crystallographic structure (receptor) were obtained from Protein Data Bank (PDB ID: DBH (Berman et al., 2000) and 2FQ1 (Drake et al., 2006), respectively). Polar hydrogens were added and Kollman charges were assigned using MGLTools (v1.5.7) (Sanner, 1999). Blind docking was performed with Achilles Blind Docking server (Bioinformatics and High Performance Computing (BIO‐HPC) Research group, n.d.), in which the whole EntB surface was made available for ligand search. Restrained docking was performed with AutoDock Vina (v1.1.2) (Eberhardt et al., 2021; Trott & Olson, ²⁰¹⁰), in which the surface of the EntB isochorismatase was interrogated for possible DHB binding sites. Conformers with the lowest binding energy from each method were selected for docking analysis using PyMOL (Schrodinger LLC, 2015). The EntB monomer (PDB ID: 2FQ1, chain A (Drake et al., 2006)) was submitted to ConSurf database (Ben Chorin et al., 2020; Goldenberg et al., ²⁰⁰⁹) for identifying the conservation of ligand‐interacting residues.

2.3. Protein expression and purification

Protein overexpression and purification were performed as reported previously (Bin & Pawelek, 2024) following transformation of expression constructs into competent AG‐1 cells following established protocols (Sambrook & Russell, 2006). Briefly, constructs containing in‐frame N‐terminal hexahistidine tagged EntB (WT and variants) and MenD harboring pCA24N plasmids (ASKA repository (Kitagawa et al., 2005)) were transformed into E. coli AG‐1 cells. Constructs containing in‐frame N‐terminal hexahistidine tagged EntC harboring pET24b plasmid were transformed into E. coli BL21(DE3) cells. Cells were grown in LB (or 2xYT for EntC) broth containing 30 μg/mL chloramphenicol (or kanamycin for EntC) at 37 °C to a final OD₆₀₀ between 0.6 and 0.8. Overexpression of recombinant proteins was induced by 0.5 mM isopropyl β‐D‐1‐thiogalactopyranoside (IPTG) overnight at 16–22°C.

Cells were lysed by ultrasonic homogenizer (Biologics) in a buffer containing 50 mM HEPES (pH 8.0), 500 mM NaCl, 5% glycerol, 10 mM imidazole, 2 mM β‐mercaptoethanol, and protease inhibitor cocktail. Cell lysate was centrifuged for 1.5 h at 47,850g, and the supernatant was collected. Clarified lysate was applied to a metal ion affinity column (Profinity IMAC resin, Bio‐Rad Laboratories) connected to a BioLogic DuoFlow FPLC system (Bio‐Rad Laboratories). Target proteins were eluted using a linear gradient of imidazole (10–400 mM imidazole) in a buffer containing 50 mM HEPES (pH 8.0), 500 mM NaCl, 5% glycerol, and 2 mM β‐mercaptoethanol. Pooled fractions containing purified proteins were dialyzed against 50 mM HEPES (pH 8.0), 150 mM NaCl, 1.0 mM tris(2‐carboxyethyl)phosphine (TCEP) and 15% glycerol. Dialyzed samples were stored at −20°C. Protein concentrations were determined by measuring absorbance at 280 nm and using molar extinction coefficients predicted from primary amino acid sequences (ε = 51,910 M⁻¹ cm⁻¹, 31,970 M⁻¹ cm⁻¹ and 107,285 M⁻¹ cm⁻¹ for EntB, EntC and MenD, respectively) (Gasteiger et al., 2005).

2.4. Fluorescence anisotropy

Reaction mixtures containing 2 μM of 2,3‐DHB and varying concentrations of protein in Assay Buffer (50 mM HEPES (pH 8.0), 150 mM NaCl and 1 mM TCEP) were preincubated at room temperature for 10 min. Fluorescence measurements (${\lambda }_{\mathrm{ex}}$ = 315 nm, ${\lambda }_{\mathrm{em}}$ = 437 nm; slit width: 10 nm) were carried out at 21°C on Varian Cary Eclipse spectrofluorometer using a 10 mm pathlength quartz cuvette (Hellma). Fluorescence anisotropy ($\gamma$) was calculated using equation $\gamma =\frac{I_{\mathit{VV}}-G\times I_{\mathit{VH}}}{I_{\mathit{VV}}+2\times G\times I_{\mathit{VH}}}$, where $I_{\mathit{VV}}$ and $I_{\mathit{VH}}$ represent fluorescence intensities obtained vertically polarized excitation (V in italic subscript) and vertically or horizontally polarized emission (V or H in italic subscript, respectively); G is an instrument‐dependent correction factor. Apparent binding affinity (K _D) values were determined by fitting the data to the one‐site total binding model (GraphPad Prism): $Y=\frac{B_{\max }\times X}{K_D+X}+\mathrm{NS}\times X+\mathrm{BG}$, where Y represents the binding signal, X represents ligand concentration, $B_{\max }$ represents maximal specific binding, NS represents nonspecific binding, BG represents background, and K _D represents the equilibrium dissociation constant.

2.5. EntB isochorismatase activity assay

Enzymological assays were carried out at 21°C using a Genesys 10 Spectrophotometer. Due to the lack of commercial availability of isochorismate, we prepared an equilibrium mixture of chorismate:isochorismate as reported previously (Bin & Pawelek, 2024) by incubating 256 or 1026 μM chorismate with 0.50 μΜ EntC in Assay Buffer containing 5 mM MgCl₂ in an ice bath for 20 min, which was expected to yield a chorismate:isochorismate ratio of approximately 3:2 (Hubrich et al., 2014). Once the reaction reached equilibrium, the chorismate:isochorismate mixture was isolated by centrifugation at 10,000g for 7 min through a 0.5 mL centrifugal filter (3K MWCO, Pall Corporation). Chorismate:isochorismate mixtures were freshly prepared on the day of use.

Reaction mixtures containing 0.10 μM EntB, 1.50 unit/mL LDH, and 350 μM NADH in Assay Buffer were preincubated at room temperature for 10 minutes. Enzymatic reactions were initiated by the addition of the chorismate:isochorismate equilibrium mixture to a final concentration range between 6.2 and 73.9 μM. Reactions were followed by measuring the decrease in NADH absorbance at 340 nm. Initial velocities were calculated by linear fitting of OD₃₄₀ decrease using the VISIONlite Rate software (Thermo Scientific). Fits of data to steady‐state enzyme kinetic models was performed using Kaleidagraph (Synergy Software).

Steady‐state kinetic assays of EntB isochorismatase were performed in the presence of 0, 100, 250 and 400 μM of 2,3‐DHB. Kinetics data were fit to the Michaelis–Menten model to determine $K_m$ and $V_{\max }$ values. In order to further characterize 2,3‐DHB inhibition behavior, Eadie–Hofstee plots of steady‐state kinetics data were generated using Kaleidagraph, and K _i values were determined by a linear replot of slope values at specific 2,3‐DHB concentrations.

2.6. Substrate channeling: MenD competition assay

Kinetic assays of EntC‐EntB‐LDH coupled reactions were carried out using similar conditions as for the EntB‐LDH coupled assay above. Excess MenD was added to the reaction mixtures (final concentration: 6.25 μM) in Assay Buffer containing 0.05 μΜ EntC, 1.25 μΜ EntB, 50 μM thiamine pyrophosphate, and 660 μM 2‐ketoglutarate (the latter two reagents being MenD co‐substrates) in the presence of 0.35 mM 2,3‐DHB. The reactions were initiated by the addition of chorismate to a final concentration of 26 μM. Initial velocities were calculated by linear regression fitting of observed OD₃₄₀ decrease using the VISIONlite Rate software (Thermo Scientific). After correction for background, isochorismatase activity in the presence of excess MenD was divided by that in the absence of MenD in order to calculate the efficiency of substrate channeling (expressed as a percentage).

2.7. Isothermal titration microcalorimetry

All ITC measurements were performed on a VP‐ITC Microcalorimeter (Microcal, Inc.). To monitor the coupled EntC–EntB enzymatic reaction by ITC, enzymes were diluted in ITC Buffer (50 mM HEPES (pH 8.0), 150 mM NaCl, 1 mM TCEP, 10 mM MgCl₂), extensively degassed, and transferred into the instrument's sample cell. The instrument syringe contained 79 μM chorismate in ITC buffer. ITC experiments were performed at a constant temperature of 30°C. Each experiment involved two syringe injections of 30 μL over a 60‐s interval with constant stirring at 300 RPM, and with a 10‐min spacing time between the injections. EntC was added to the sample cell at a final concentration of 313 nM in the absence or presence of EntB (wild‐type or variant as appropriate) (concentration range: 162–2587 nM). Integrated reaction heats were quantitated using Microcal Origin (Microcal, Inc.). Integrated heats of enzymatic reactions were corrected for the heat of injection (30 μL ITC buffer injected into ITC buffer), as well as the heat of dilution of chorismate into ITC buffer minus protein.

3. RESULTS AND DISCUSSION

Our initial observation of 2,3‐DHB interaction with E. coli EntB arose from a broader study we performed in which we investigated the ability of various DHB isomers to quench intrinsic fluorescence of EntB and EntE, and found that 2,3‐DHB was able to specifically quench EntB fluorescence unlike 3,5‐DHB and 2,5‐DHB (Khalil & Pawelek, 2009). Given our recent finding that EntC and EntB interact (Ouellette et al., 2022) to form a dynamic channeling system (Bin & Pawelek, 2024), we decided to revisit and expand on this observation to see if 2,3‐DHB might also play a functional role in the EntCB coupled reaction.

3.1. Computational modeling of 2,3‐DHB binding at the EntB active site

To identify possible 2,3‐binding sites on the surface of EntB, we performed automated docking experiments, taking both a blind docking approach as well as a restrained docking approach. The blind docking simulation using 2,3‐DHB as a ligand and dimeric EntB as a receptor produced 24 energetically favorable docked ensembles. The ensembles were ranked by the binding energy, which ranged from −7.0 to −3.6 kcal/mol. The most energetically favorable pose of 2,3‐DHB was located at the active site of EntB (Figure 2). Moreover, residues involved in the predicted EntB‐DHB interaction overlapped with six out of 10 residues reported by Drake et al. (2006) as being catalytically important. In addition to blind docking simulations, a restrained docking simulation was performed that produced nine energetically favorable models (binding energy range: −6.9 to −6.0 kcal/mol); the pose of the top‐ranked model from the restrained docking approach was superimposable with the pose of the top‐ranked model from the blind docking approach (RMSD = 0.009 Å) (Figure 2, cyan and orange binding poses).

FIGURE 2.

Automated docking simulation of EntB binding to 2,3‐DHB. EntB shown in cartoon presentation (helices = coils; strands = arrows); ligand (2,3‐DHB) and EntB residues in contact with ligand are shown as sticks with carbon atoms colored as follows: Green = EntB catalytic residues; dark blue = EntB residues in contact with bound ligand; orange = the top‐ranked DHB binding pose from blind docking; cyan = top‐ranked DHB binding pose from restrained docking (RMSD = 0.009 Å).

3.2. Equilibrium‐binding of 2,3‐DHB to EntB: Fluorescence anisotropy

We previously reported the binding of DHB isomers to EntB via fluorescence spectroscopy, in which the EntB intrinsic fluorescence was found to be most strongly quenched by 2,3‐DHB in comparison with other DHB analogues (2,5‐dihydroxybenzoic acid (2,5‐DHB) and 3,5‐dihydroxybenzoic acid (3,5‐DHB)) (Khalil & Pawelek, 2009). Since 2,3‐DHB is fluorescent (${\lambda }_{\mathrm{ex}}$ = 315 nm, ${\lambda }_{\mathrm{em}}$ = 437 nm), here we measured changes in 2,3‐DHB fluorescence anisotropy to better understand equilibrium‐binding characteristics of 2,3‐DHB when interacting with relevant target proteins (Figure 3). We found that 2,3‐DHB bound to EntE with relatively high affinity (apparent K _D = 0.54 μM), consistent with 2,3‐DHB being a natural substrate of EntE (Figure 3a). We further found that 2,3‐DHB binds specifically to EntB, exhibiting binding behavior consistent with the single‐site binding model (apparent K _D = 8.95 μM) (Figure 3b). Additionally, we previously reported (Bin & Pawelek, 2024) that EntB residue R196 is essential for isochorismate channeling between EntC and EntB as well as electrostatically guiding isochorismate to the EntB active site. Fluorescence anisotropy experiments further revealed that 2,3‐DHB binds to the EntB variant R196A with much lower affinity than wild‐type EntB (Figure 3c; apparent K _D not determinable due to insufficient approach to saturation). The lowered affinity of 2,3‐DHB for this variant, previously shown to be deficient in EntCB isochorismatase channeling, is consistent with our previous observation that EntB R196 is also involved in electrostatically guiding isochorismate to the isochorismatase active site (Bin & Pawelek, 2024).

FIGURE 3.

Fluorescence anisotropy binding assays. (a) EntE, (b) wild‐type EntB, (c) EntB R196A. In all cases, purified protein was titrated into 2.00 μM of 2,3‐DHB and fluorescence anisotropy was measured at each protein concentration (${\lambda }_{\mathrm{ex}}$ = 315 nm, ${\lambda }_{\mathrm{em}}$ = 437 nm). Data are shown as the mean of triplicate independent measurements; error bars represent standard deviations from the mean. Apparent K _D values determined from fits of the data to the single‐site binding model; apparent K _D for 2,3‐DHB binding to EntB R196A not determinable due to insufficient approach to saturation over concentration range.

3.3. EntB isochorismatase: Steady‐state kinetics

EntB isochorismatase activity was measured in a coupled assay with excess LDH as reported previously (Bin & Pawelek, 2024). In this assay, isochorismate is converted to 2,3‐dihydro‐2,3‐dihydroxybenzoate (2,3‐diDHB) and pyruvate by EntB isochorismatase activity; LDH‐catalyzed conversion of pyruvate to lactate is then monitored by loss of OD₃₄₀ following oxidation of NADH to NAD⁺. The Hill coefficient from an IC₅₀ dose–response curve of EntB isochorismatase inhibition by 2,3‐DHB was 1.16 (Figure S1), consistent with a DHB:EntB binding stoichiometry of 1.0 (Prinz, 2010). Coupled assay kinetics data for the conversion of isochorismate to 2,3‐diDHB were fit to the Michaelis–Menten steady‐state kinetics model, resulting in an apparent Michaelis–M enten constant (K _{m app}) equal to 9.2 μM (Table 1, Row 2, Column 2), which is similar to a published K _m value for EntB isochorismatase (K _m = 14.7 μM), as determined using a different experimental approach (Rusnak et al., 1990). Given this similarity, we concluded that our coupled assay was suitable for investigating possible inhibition of EntB isochorismatase activity by 2,3‐DHB in the absence of EntC.

TABLE 1.

Steady‐state kinetic parameters of wild‐type EntB isochorismatase in the absence and presence of 2,3‐DHB.

[2,3‐DHB] (μM)	Isochorismate apparent K m (μM)	V max (μM min−1)
0	9.21 ± 0.97	4.91 ± 0.22
100	16.24 ± 1.57	5.25 ± 0.17
250	22.12 ± 3.25	4.99 ± 0.52
400	32.55 ± 0.23	5.33 ± 0.05

Note: Enzyme kinetics data were obtained using the EntB‐LDH coupled enzyme assay (0.1 μM EntB) initiated with a 3:2 equilibrium mixture of chorismate and isochorismate.

3.4. Inhibition of EntB isochorismatase by 2,3‐DHB

We performed a series of steady‐state kinetic assays in which EntC, EntB, and LDH activities were coupled (Bin & Pawelek, 2024) to investigate the effect of increasing concentrations of 2,3‐DHB on overall coupled enzyme activity. Our kinetics data showed that the EntB K _m value for isochorismate decreased as a function of 2,3‐DHB concentration (Table 1, Column 2), whereas V _max values were unaffected by the presence of 2,3‐DHB (Table 1, column 3). Kinetics data were further analyzed using an Eadie–Hofstee plot. Linear fits of the data were observed to approximately converge at the ordinate, indicative of competitive inhibition (Figure 4). A replot of the slopes obtained from the Eadie–Hofstee linear fits (i.e., the apparent K _m values for isochorismate) as a function of DHB concentration resulted in an apparent K _i value of 200 μM (Figure 4, inset). This K _i value is similar to the intracellular concentration of 2,3‐DHB in exponentially growing E. coli cells (~140 μM) (Bennett et al., 2009), supporting a biological role for 2,3‐DHB in the inhibition of EntB isochorismatase.

FIGURE 4.

Eadie–Hofstee plot of EntB‐LDH coupled assay kinetics in the presence and absence of 2,3‐DHB. Circles: 0 μM 2,3‐DHB, diamonds: 100 μM 2,3‐DHB, squares: 250 μM 2,3‐DHB, triangles: 400 μM 2,3‐DHB. Data shown with the mean and standard deviation of triplicate independent measurements. Inset: Replot of linear slope values as a function of 2,3‐DHB concentration. The K _i value was determined from the slope of the linear replot.

We previously reported partial channeling of isochorismate in an EntC‐EntB‐LDH coupled assay in which approximately 16% residual isochorismatase activity was observed in the presence of a competing isochorismate‐utilizing enzyme, MenD (Bin & Pawelek, 2024). Here, we repeated our channeling experiment, but in the presence of 2,3‐DHB. Residual EntB isochorismatase activity in the presence of excess MenD, a measure of substrate channeling efficiency, was observed to decrease to approximately 5% in the presence of 2,3‐DHB (Figure 5), reflecting an overall 70% decrease in isochorismate channeling efficiency in the presence of 350 μM 2,3‐DHB. Consistent with our competitive inhibition data, we conclude that 2,3‐DHB binds at the active site of EntB isochorismatase, and in doing so obstructs the continuous channeling surface between EntC and EntB active sites (Bin & Pawelek, 2024).

FIGURE 5.

The effect of 2,3‐DHB on the EntC–EntB isochorismate channeling. Residual EntB Activity = EntB isochorismate activity in the presence of 350 μM 2,3‐DHB and MenD (0–6.25 μM). Data were measured in triplicate. Standard deviations from the mean are indicated with error bars. Dashed line indicates 16% EntC–EntB isochorismate channeling measured previously reported (Imperi & Visca, 2013) under identical conditions but in the absence of 2,3‐DHB.

3.5. ITC measurements of EntC–EntB reaction and inhibition by DHB

Isothermal titration microcalorimetry is a technique in which heat evolved or absorbed by a reaction is measured, allowing for calculation of thermodynamic parameters (ΔG, ΔH, ΔS). Typically, ITC is used for studying binding reactions given its utility in determining K _D and stoichiometry values of a particular binding system. In addition, ITC has been successfully used to follow steady‐state enzyme kinetics (Siddiqui et al., 2022; Wang et al., ²⁰²⁰). We, therefore, decided to employ ITC to follow the EntC–EntB coupled reaction in real time without the need for a coupled reporter enzyme like LDH. We performed recurrent single‐injection assays (Wang et al., 2020) in which the EntC substrate (chorismate) was injected into a sample cell containing dilute EntC with or without EntB. Two injections of 79 μM chorismate into a sample cell containing 313 nM EntC in ITC buffer resulted in very low heat evolution compared with a control injection lacking chorismate, suggesting that the EntC‐catalyzed isomerization of chorismate to isochorismate has a relatively low enthalpy change. Repetition of the experiment, but with the addition of EntB to the sample cell, resulted in a large increase in heat evolution in a manner dependent on the concentration of added EntB (Figure 6a). The rapid return to baseline as a function of EntB concentration for each injection supports our prior report of EntC–EntB substrate channeling (Bin & Pawelek, 2024). Given the complexity of the EntC–EntB coupled reaction (i.e., dynamic channeling following EntC–EntB pairwise interaction, each having its own heat signature in addition to those of the intrinsic EntC and EntB enzymatic reactions), deconvolution of observed heats into discrete kinetic rate constants was not possible. However, integration of thermal spikes following chorismate injection allowed for relative comparison of EntCB activity as a function of increasing EntB concentration (Figure 6b). Overall heat evolution was observed to rapidly increase up to an EntB:EntC molar ratio of 1:1, followed by a more gradual increase going from 2:1 to 8:1 EntB:EntC. The observed gradual increase is likely due to approach to saturation of EntCB complex formation at higher EntB concentrations, since dynamic channeling requires transient interaction of EntC and EntB (Bin & Pawelek, 2024). We also performed the EntC–EntB ITC experiment in the absence of Mg²⁺. In this case, only the heat of injection was observed, confirming that the observed heats were indeed due to magnesium‐dependent EntC enzymatic activity (Figure 6c). Furthermore, it has previously been reported that the EntC‐catalyzed reaction is highly reversible (Liu et al., 1990), and that EntB may assist in metabolic flux in the direction of 2,3‐DHB formation by rapidly converting isochorismate to 2,3‐diDHB thus preventing its EntC‐catalyzed conversion back to chorismate. Our ITC results support these observations. Since we could use ITC to interrogate the effect of EntB on the EntC reaction, we also investigated the effect of 2,3‐DHB addition to the sample cell while EntB was present. We found that heat evolution decreased in a DHB‐dependent manner (Figure 6d) consistent with our analysis of EntCB steady‐state kinetics data in which the competitive inhibition of EntB by 2,3‐DHB was found to have an apparent K _i value of approximately 200 μM.

FIGURE 6.

Isothermal titration microcalorimetry of the EntC‐catalyzed reaction. (a) Recurrent single‐injection assays of EntC enzymatic activity in the absence and presence of EntB. Raw thermograms showing heats evolved upon injection of chorismate into a reference cell containing 313 nM EntC and increasing concentrations of EntB (left panel: 0 nM EntB, middle panel: 323 nM EntB, right panel: 1293 nM EntB). All panels rendered to the same scale. (b) Corrected integrated heats of the EntC‐catalyzed reaction as a function of EntB concentration. (c) Raw thermograms showing magnesium dependence of the EntB‐driven EntC reaction (313 nM EntC, 1293 nM EntB): Blue: 10 mM MgCl₂, red: 0 mM MgCl₂ + 10 mM EDTA. Thermograms arbitrarily offset along the ordinate for the purpose of comparison. (d) Raw thermograms showing the effect of 2,3‐DHB on the EntB‐driven EntC reaction (313 nM EntC, 1293 nM EntB): Blue: 0 μM DHB, red: 250 μM DHB, black: 500 μM DHB. Thermograms arbitrarily offset along the ordinate for the purpose of comparison.

4. CONCLUSIONS

Although several chorismate analogs have been reported to inhibit EntB isochorismatase activity (Hubrich et al., 2013; Rusnak et al., ¹⁹⁹⁰), this study presents the first evidence that 2,3‐DHB can function as a competitive inhibitor of EntB isochorismatase. To further explore the nature of 2,3‐DHB binding to EntB, which we had first observed as 2,3‐DHB‐dependent quenching of intrinsic EntB fluorescence (Khalil & Pawelek, 2009), we employed automated docking simulations and equilibrium‐binding experiments. Both molecular docking simulations and steady‐state enzyme kinetics studies demonstrated that 2,3‐DHB is a competitive inhibitor of EntB isochorismatase. Furthermore, ITC experiments revealed that the rate of conversion of chorismate to 2,3‐diDHB increased in an EntB‐dependent manner, supporting our previous observation of EntCB channeling of isochorismate. Taken together, these in vitro observations suggest that 2,3‐DHB may play a biological role as a negative feedback regulator of enterobactin biosynthesis via direct inhibition of EntB following accumulation of sufficient intracellular iron. Given the highly reversible nature of EntC activity, inhibition of EntB isochorismatase therefore would target the committed step of the enterobactin biosynthetic pathway. Disruption of isochorismate channeling by 2,3‐DHB would result in non‐productive diffusion of the labile intermediate into the bulk cytosol or re‐routing to other metabolic fates. Similar feedback regulation has also been found in the enzyme complex that catalyzes the first two sequential reactions of tryptophan biosynthesis in E. coli, anthranilate synthetase (TrpE) and phosphoribosyl transferase (TrpD), where tryptophan acts on TrpE while cooperatively inhibiting TrpD in its complex formed with TrpE. On the other hand, the regulation of TrpE is allosteric inhibition instead of competitive inhibition (Naz et al., 2023; Pabst et al., ¹⁹⁷³). Given the promiscuity of EntB binding to other enzymes in the enterobactin biosynthetic pathway (EntC, EntA, EntE, and EntF), we are now interested in understanding this interplay between biosynthesis and its regulation in the context of larger multienzyme assemblies (i.e., EntCBAE). Moreover, these results reveal that EntB may serve as an excellent target for disruption of enterobactin‐mediated iron uptake in a therapeutic context.

AUTHOR CONTRIBUTIONS

Xue Bin: Investigation; writing – original draft; methodology; validation; visualization; writing – review and editing. Peter D. Pawelek: Conceptualization; investigation; funding acquisition; writing – original draft; methodology; validation; visualization; writing – review and editing; supervision; resources.

Supporting information

FIGURE S1. IC₅₀ analysis of EntB isochorismatase inhibition by 2,3‐DHB. Blue: wild‐type EntB; Red: EntB R196A. Non‐linear fitting resulted in IC₅₀ values of 360 μM and 692 μM for EntB wild‐type and R196A, respectively. Data shown are mean values and standard deviations of triplicate independent measurements.

Bin X, Pawelek PD. The enterobactin biosynthetic intermediate 2,3‐dihydroxybenzoic acid is a competitive inhibitor of the Escherichia coli isochorismatase EntB . Protein Science. 2025;34(6):e70160. 10.1002/pro.70160

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.