Eliminating competition: Characterizing and eliminating competitive

Aug 3, 2018 - though other metals show slower association/dissociation kinetics.11, 19. DAHP oxime is a transition state mimic inhibitor of DAHPS, wit...
0 downloads 0 Views 2MB Size
Download PDF
Article Cite This: Biochemistry XXXX, XXX, XXX−XXX

pubs.acs.org/biochemistry

Eliminating Competition: Characterizing and Eliminating Competitive Binding at Separate Sites between DAHP Synthase’s Essential Metal Ion and the Inhibitor DAHP Oxime Maren Heimhalt,† Shan Jiang,‡ and Paul J. Berti*,†,‡ †

Department of Chemistry & Chemical Biology and ‡Department of Biochemistry & Biomedical Sciences, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4M1, Canada

Downloaded via UNIV OF LEICESTER on November 14, 2018 at 13:12:04 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: 3-Deoxy-D-arabinoheptulosonate 7-phosphate (DAHP) oxime is a transition state mimic inhibitor of bacterial DAHP synthase, with Ki = 1.5 μM and a residence time of tR = 83 min. Unexpectedly, DAHP oxime inhibition is competitive with respect to the essential metal ion, Mn2+, even though the inhibitor and metal ion do not occupy the same physical space in the active site. This is problematic because DAHP synthase is activated by multiple divalent metal cations, some of which have significant intracellular concentrations and some of which dissociate slowly. The nature of DAHP oxime’s competition with the metal ion was investigated. Inhibition shifted from metal-competitive at physiological pH to metal-noncompetitive at pH > 8.7 in response to deprotonation of the Cys61 side chain. The modes of inhibition of DAHP synthase mutants and inhibitor fragments demonstrated that metal-competitive inhibition arose from interactions between Mn2+, DAHP oxime’s O4 hydroxyl group, and the Cys61 and Asp326 side chains. The majority of potent DAHP synthase inhibitors in the literature possess a 4-hydroxyl group. Removing it could avoid metal-competitive inhibition and avoid them being outcompeted by metal ions in vivo.

T

occupy the same physical space.18 Understanding this competition is critical for improving DAHPS inhibitors because the abundance of metals in living cells can allow them to outcompete inhibitor binding.23 In addition, some metals dissociate very slowly,12,17,18 which could block inhibitor binding. Most DAHPS inhibitors have been designed to mimic the THI or the presumed cationic intermediate/ transition state for THI formation.2−6 Like DAHP oxime, they may be subject to metal competition, so it is important to understand the nature of metal-competitive inhibitor binding. We report here that DAHP oxime’s mode of inhibition changed from metal-competitive at physiological pH to metalnoncompetitive at high pH due to ionization of the Cys61 side chain. The modes of inhibition of enzyme mutants and fragment-based inhibitors demonstrated that metal-competitive inhibition arose from interactions between Mn2+, DAHP oxime’s O4 hydroxyl group, and the C61 and D326 side chains. Removing these interactions made inhibition metal-noncompetitive.

he bacterial enzyme 3-deoxy-D-arabinoheptulosonate 7phosphate synthase (DAHPS) is a member of the αcarboxyketose synthase superfamily1 and a potential antimicrobial target.2−7 It catalyzes an aldol-like reaction of phosphoenolpyruvate (PEP) and erythrose 4-phosphate (E4P), which passes through a tetrahedral intermediate (THI) on the way to forming DAHP and inorganic phosphate (Pi) (Scheme 1).8,9 DAHPS requires a divalent metal ion activator,10 and several metals have been proposed as the physiological activator, including Fe2+,11−14 Co2+,15,16 and Cu2+.17 Many other metal ions activate DAHPS, with Mn2+ commonly being used because of its high activity and rapid equilibrium kinetics.11,18,19 DAHPS follows a sequential ordered ter ter kinetic mechanism, with the metal ion binding first, followed by PEP, and then E4P.18,20,21 Escherichia coli DAHPS(Phe), the phenylalanine-sensitive isozyme, has a rapid equilibrium kinetic mechanism with Mn2+ as the activator,18 though other metals show slower association/dissociation kinetics.11,19 DAHP oxime is a transition state mimic inhibitor of DAHPS, with Ki = 1.5 μM and a residence time, tR, of 83 min.18,22 The oxime functional group structurally mimics the interaction of the THI’s/PEP’s phosphate group with the active site,18 and it functionally mimics a transition state, presumably Pi departure during THI breakdown.22 DAHP oxime binding is necessarily competitive with respect to PEP and E4P, since they occupy the same physical space in the active site. Surprisingly, it is also competitive with respect to Mn2+ even though they do not © XXXX American Chemical Society

Received: August 8, 2018 Revised: October 31, 2018 Published: November 6, 2018 A

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry Scheme 1. DAHPS Reaction

■

changing the fitted values by (presumably) reducing the covariance between individual KM values. The pH profiles of different kinetic parameters were fitted empirically, since, especially for KM values, there is no expectation of a particular shape of curve. KM,Mn decreased monotonically at high pH and was fitted to a single pKa with an exponent of 3 (eq 2)

MATERIALS AND METHODS General. DAHPSH6, the N-terminally His6-tagged Phesensitive isozyme of E. coli DAHPS in the aroG:pCA24N expression plasmid,24 and mutant enzymes were expressed, purified, and demetalated as described previously.22 Erythrose 4phosphate (E4P) and DAHP oxime were synthesized as described previously.18,25 Other reagents and solvents were purchased from Sigma-Aldrich or Bioshop Canada (Burlington, ON) and were used without further purification. Pyruvate Oxime and Glyoxylate Oxime. Pyruvate oxime was synthesized by combining 5 mmol of sodium pyruvate and 5 mmol of hydroxylamine hydrate (50 wt % in H2O) in 1 mL of water, adjusting to pH 6.3, and then stirring at room temperature for 3 h. Glyoxylate oxime was synthesized using the same procedure, using 5 mmol of glyoxylic acid monohydrate. Pyruvate oxime: HRMS (m/z): [M − 1]− calcd for C3H4NO3, 102.0197; found, 102.0199. 1H NMR (200 MHz, D2O): δ 1.99 (3H, s, H3); 13C NMR (125 MHz, D2O): δ 11.20 (C3), δ 156.16 (C2) δ 171.06 (C1). Glyoxylate oxime: HRMS (m/z): [M − 1]− calcd for C2H3NO3, 88.0040; found, 88.0039. 1H NMR (200 MHz, D2O) H2 δ 7.512 (1H, s, H2); 13C NMR (50 MHz, D2O): δ 147.63 (C2) δ 169.39 (C1). Rate Assays. Initial velocities were measured using the Malachite green/ammonium molybdate colorimetric assay for Pi production.18,26,27 All solutions except the enzyme and MnCl2 were treated before use with Chelex 100 to remove metals. At pH 7, assay conditions were as described previously, with two substrate concentrations held constant while the third was varied.18 The kinetics buffer contained 50 mM buffer component, 100 mM KCl, and 0.1 mM tris(2-carboxyethyl) phosphine (TCEP) at 25 °C. The buffer component was the following: pH 7−8, potassium 4-(2-hydroxyethyl)-1-piperazineethanesulfonate (K-HEPES); pH 8−9, potassium N-cyclohexyl-2-aminoethanesulfonate (K-CHES); pH 10−10.5, potassium N-cyclohexyl-3-aminopropanesulfonate (K-CAPS). Substrate concentrations were adjusted as needed to obtain reliable reaction rates, using 1−10 μM MnCl2, 0.1−5 mM PEP, 0.1−0.5 mM E4P, and 25−100 nM DAHPSH6. At pH ≥ 10, manganese hydroxide precipitated at extended reaction times, so MnCl2 and DAHPSH6 were preincubated in storage buffer at pH 7.5 and then the reaction was started by adding them together to the other substrates in the high pH buffer. Kinetic Parameters and pH Profiles. Mn2+ is an essential activator of DAHPS, but it can be treated as a substrate in initial velocity calculations.18,28 Kinetic parameters were determined by fitting initial velocities to eq 1 for a rapid equilibrium sequential ordered ter ter kinetic mechanism:28 v0 = [E]0 1+

KM,Mn′ =

+

[Mn][PEP] KM,MnKM,PEP

+

[Mn][PEP][E4P] KM,MnKM,PEPKM,E4P

103 × (pKa − pH) + 1

+ KM,Mn(high) (2)

where KM,Mn′ is the pH-dependent value of KM,Mn and KM,Mn(low) and KM,Mn(high) are its values at low and high pH. KM,PEP′ had a U-shaped pH profile and was fitted to eq 3 KM,PEP′ =

KM,PEP(low) × 10(pKa1− pH) 1 + 10(pKa1− pH) + +

KM,PEP(mid) × 10(pH − pKa1) 10(2 × pH − pKa1− pKa2) + 10(pH − pKa1) + 1 KM,PEP(high) × 10(pH − pKa2) 1 + 10(pH − pKa2)

(3)

where KM,PEP′ is the pH-dependent value of KM,PEP, KM,PEP(low) and KM,PEP(high) are its values at the acidic and basic limits, KM,PEP(mid) is its minimum value, and pKa1 and pKa2 are the acid- and base-limb pKa values, respectively. The pH profiles for kcat′, KM,E4P′, and kcat/(KM,Mn, KM,PEP, KM,E4P) ′ were bell-shaped (eq 4) X′ =

X(max) × 10(pH − pKa1) 10(2 × pH − pKa1− pKa2) + 10(pH − pKa1) + 1

(4)

where X′ is the pH-dependent value of each parameter, X(max) is its maximum value, and pKa1 and pKa2 are acid- and base-limb pKa values, respectively. pH Dependence of DAHP Oxime Inhibition. Substrate concentrations were held constant at each pH, while DAHP oxime was varied from 0−800 μM. Equation 5 takes into account the pH-dependent change in the proportion of metalcompetitive versus -noncompetitive inhibition and the Ki values for both modes of inhibition. Equation 5 contains rate terms for metal-competitive (vcomp/[E]0) and -noncompetitive (vnoncomp/ [E]0) binding:

kcat[Mn][PEP][E4P] KM,MnKM,PEPKM,E4P [Mn] KM,Mn

KM,Mn(low) × 103 × (pKa − pH)

ij ij 10 pH − pKa v0 10 pH − pKa yzz vcomp = jjj1 − pH − pK + jjj pH − pK zz × a a [E]0 [E]0 + 1{ + 10 k k 10 vnoncomp v × + offset [E]0 [E]0

(1)

Equation 1 was used to determine kcat directly. The same equation was used to fit the specificity constant kcat/ (KM,MnKM,PEPKM,E4P) as a single parameter, as well as the individual KM values.18 This lowers the standard errors without B

yz zz z 1{

(5)

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry Scheme 2. Kinetic Mechanism of DAHP Oxime Inhibition (eq 5)a

a

EH and E are the low and high pH forms of the enzyme, respectively. At pH < 7.7, DAHP oxime inhibition is competitive with respect to Mn2+ (black), while it is non-competitive at pH > 9.7 (grey). Ka refers to the EH ⇌ E equilibrium constant, not the individual pKas of each pHdependent kinetic parameter.

Figure 1. pH dependence of DAHPSH6’s kinetic parameters: (a) kcat/KM,MnKM,PEPKM,E4P, the specificity constant, (b) kcat, (c) (top) KM,Mn, (middle) KM,PEP, (bottom) KM,E4P.

vcomp [E]0

vnoncomp v0 v = + offset [E]0 [E]0 [E]0

= kcat ′ [Mn][PEP][E4P] KM,Mn ′ KM,PEP ′ KM,E4P′

1+

vnoncomp [E]0 jij1 + j k

[I] K i ,comp

+

[Mn] KM,Mn′

+

[Mn][PEP] KM,Mn ′ KM,PEP′

+

Ki,comp and Ki,noncomp are the metal-competitive and -noncompetitive inhibition constants, respectively; the kinetic parameters (kcat′, KM,Mn′, etc.) are the pH-dependent parameters from eqs 2−4 and are held constant. A voffset/[E]0 term is included to account for DAHP oxime’s partial inhibition, with ∼15% residual activity at high inhibitor concentrations.18 Ki values for the fragment pairs were fitted using eq 6 or 7. ITC Titrations. Isothermal titration calorimetry (ITC) was performed using a NanoITC calorimeter (TA Instruments, Delaware, MD). Titrations of DAHPSH6·Mn2+ with DAHP oxime at pH 9 were performed by exchanging the enzyme immediately before use into 20 mM K-CHES, pH 9.0, and 0.1 mM TCEP and then concentrating it to 100 μM by centrifugal ultrafiltration. DAHP oxime (400 μM) was prepared in the ultrafiltrate, while MnCl2 was prepared in water. After degassing

[Mn][PEP][E4P] KM,Mn ′ KM,PEP ′ KM,E4P′

= [I] zyzjij1 K i ,noncomp zj

{k

kcat ′ [Mn][PEP][E4P] KM,Mn ′ KM,PEP ′ KM,E4P′

+

y [Mn] z z KM,Mn′ z

vcomp v0 v = + offset [E]0 [E]0 [E]0

{

+

[Mn][PEP] KM,Mn ′ KM,PEP′

+

(7)

[Mn][PEP][E4P] KM,Mn ′ KM,PEP ′ KM,E4P′

(6) C

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry Table 1. DAHPSH6’s pH Dependencea kinetic parameter Ki′ b

kcat/(KM,MnKM,PEPKM,A5P)′ c

kcat′ c

KM,Mn′ d

KM,PEP′ e

KM,A5P′ c

pH dependence

value

pKa Ki,comp Ki,noncomp pKa1 pKa2 kcat/(KM,MnKM,PEPKM,A5P)(max) pKa1 pKa2 kcat(max) pKa KM,Mn(low) KM,Mn(high) pKa1 pKa2 KM,PEP(low) KM,PEP(mid) KM,PEP(high) pKa1 pKa2 KM,A5P(max)

8.7 ± 0.1 3.7 ± 0.6 μM 6.6 ± 0.9 μM 7.0 ± 0.2 9.8 ± 0.2 (2.7 ± 0.2) × 1015 M−3 s−1 6.4 ± 0.2 10.6 ± 0.2 20 ± 1 s−1 8.2 ± 0.1 8.9 ± 0.4 μM 0.3 ± 0.2 μM 5.1 ± 0.7 10.4 ± 0.3 0.01 ± 0.02 M 15 ± 4 μMf 0.01 ± 0.01 M 7.9 ± 0.6 8.9 ± 0.5 170 ± 90 μM

a Initial velocities at each pH were fitted to eq 1; then, each kinetic parameter’s pH dependence was fitted to eqs 2−4. They were then used as constants in eq 5 to fit the pH dependence of inhibition. bFitted to eq 5. cFitted to eq 4. dFitted to eq 2. eFitted to eq 3. fFixed at the experimental value at pH 8.

all solutions for 15 min, 400 μM MnCl2 was added to DAHPSH6 and 170 μL of DAHPSH6·Mn2+ was equilibrated with stirring at 300 rpm at 20 °C before titration with ultrafiltrate (blank) or DAHP oxime. The first injection was 0.48 μL, followed by 18 × 2.5 μL injections. Kd values were fitted using an independent one-site binding model in the NanoAnalyze software (TA Instruments). Titrations of DAHP oxime binding to the D326A mutant were performed similarly but using 800 μM DAHP oxime to titrate 120 μM D326A in 20 mM Tris−Cl, pH 7.0, 0.1 mM TCEP, and 800 μM MnCl2. Mode of Inhibition from Eadie−Hofstee Plots. Eadie− Hofstee plots were used to determine the mode of inhibition with respect to the metal ion. Initial velocities were measured using varying inhibitor and metal ion concentrations and fixed concentrations of PEP and E4P in kinetics buffer. For C61A, the concentrations were 5 μM to 10 mM MnCl2, 0.25−4 mM DAHP oxime, 0.6 mM PEP, 1 mM E4P, and 2 μM enzyme. For H268A, the concentrations were 5 μM to 10 mM MnCl2, 0.04− 1 mM DAHP oxime, 0.5 mM PEP, 0.05 mM E4P, and 0.4 μM enzyme. For the pyruvate oxime + glycerol 3-phosphate (Gro3P) fragment pair, the concentrations were 2−1000 μM MnCl2, 0.8−12.5 mM pyruvate oxime, 4 mM Gro3P, 100 μM PEP, 100 μM E4P, and 25 nM DAHPSH6. Pyruvate oxime’s Ki,noncomp value was fitted to eq 7. For the glyoxylate oxime + erythritol 4phosphate (Ero4P) fragment pair, the concentrations were 50− 1000 μM MnCl2, 5−65 mM glyoxylate oxime, 8 mM Ero4P, 250 μM PEP, 125 μM E4P, and 25 nM DAHPSH6. Glyoxylate oxime’s Ki,comp value was fitted to eq 6.

■

kcat′ and KM,E4P′ all had bell-shaped pH profiles, while KM,PEP′ was U-shaped, and KM,Mn′ decreased monotonically. Because DAHPSH6 was unstable at pH < 6.5, the curves at pH ≤ 7 were often defined by a single point, so the fitted values in this region were generally not well-defined. The pH dependence of kcat/ (KM,Mn, KM,PEP, KM,E4P)′, with its optimum pH of 8.4, was dominated by KM′, since kcat′ was only weakly pH dependent. The 21-fold decrease in KM,Mn′ from pH 7 to 10 suggested that deprotonation of a metal-binding side chain increased its metal affinity. Of the metal-binding side chains, C61 was the most likely candidate, since the pKa value of 8.2 was similar to an unperturbed Cys side chain, 8.3.18,29 The other metal-binding side chainsH268, D326, and E302are likely already deprotonated at neutral pH. The C61 side chain’s effect on KM,Mn′ was tested with the C61A mutant. C61A’s KM,Mn′ value increased 2.2-fold from pH 7 to 10, from 90 ± 10 to 200 ± 50 μM, in contrast to the 21-fold decrease in DAHPSH6. Thus, C61 had a significant effect on the pH dependence of KM,Mn, though other factors must also have been contributing to the steep pH dependence. pH-Dependent Mode of Inhibition. Previously, ITC titrations showed that DAHP oxime binding was competitive with Mn2+ at pH 7, with no detectable binding to the DAHPS· Mn2+ complex.18 In the present study, ITC titrations at pH 9 showed that binding was no longer competitive, with Kd = 4.1 ± 0.7 μM for DAHP oxime binding to the DAHPSH6·Mn2+ complex (Figure 2). Initial velocities were then fitted to a kinetic model that accommodated a pH-dependent change from metal-competitive to -noncompetitive inhibition (eq 5, Figure 3, Table 1). The fit was excellent across the pH range, demonstrating a change in the mode of inhibition with pKa = 8.7 ± 0.1. The Ki values for each mode of inhibition were similar, with Ki,comp = 3.7 ± 0.6 μM and Ki,noncomp = 6.6 ± 0.9 μM. The Ki,comp value at low pH was modestly higher than the Ki = 1.5 μM measured directly at pH 7.0.18 The latter value is likely correct, since the fitted curve at

RESULTS

pH Dependence. The pH dependence of DAHPSH6’s kinetic parameters was measured in the pH range 7−10 (Figure 1, Table 1). Each was fitted empirically, since there is no expectation of a particular curve shape, especially for the KM values. The specificity constant kcat/(KM,Mn, KM,PEP, KM,E4P)′ plus D

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 2. ITC titration of the DAHPSH6·Mn2+ complex with DAHP oxime at pH 9. (top) Raw ITC trace, (bottom) integrated heat fitted to a one-site binding model, Kd = 4.1 ± 0.7 μM. The thin black line is the residual (observed − fitted), offset by −100 μJ.

pH 7 using eq 5 was slightly but consistently above the data points (Figure 3), indicating that, while the model provided an excellent overall fit to the experimental data, it overestimated Ki,comp at low pH. The value of Ki,noncomp, 6.6 μM, was consistent with the Kd = 4.1 μM value for DAHP oxime’s binding to DAHPSH6·Mn2+ by ITC titration under somewhat different conditions. It is possible that Ki,comp and Ki,noncomp are themselves pH dependent; however, the pH ranges in which one mode is dominant were too narrow for this to be determined for the wild type enzyme. In the pH range of pKa ± 1, both modes contribute significantly to the initial velocities, meaning that purely competitive inhibition only occurred in the range pH 7.0−7.7 and purely noncompetitive in the range pH 9.7−10.0. DAHPSH6 Mutants. Residues C61, H268, and D326 were investigated, as they can make contact with both the metal and DAHP oxime (Figure 4).18,30 DAHP oxime binding to the D326A·Mn2+ complex was measurable by ITC titration, with Kd = 14 ± 5 μM (Figure 5a). This was similar to the Ki = 18 μM value reported previously under slightly different conditions.22 Binding was too weak with C61A and H268A to be followed by ITC, so the modes of inhibition were determined using Eadie− Hofstee plots (v0 versus v0/[S]). With C61A, parallel lines in the Eadie−Hofstee plot revealed that DAHP oxime binding was noncompetitive with respect to Mn2+ (Figure 5b). With H268A, the lines converged toward the y-axis, indicating competitive binding (Figure 5c). With classical competitive inhibitors, the lines converge at the y-axis; however, when there is a residual rate at high inhibitor concentrations due to partial inhibition (voffset), the lines converge toward, but not at, the y-axis (see the Supporting Information). Fragment-Based Inhibitors. The DAHPS crystal structures indicate that the metal ion and DAHP oxime’s O4 hydroxyl group should be in close proximity to each other. Two pairs of

Figure 3. pH-dependent DAHP oxime inhibition. All of the initial velocity data from pH 7−10 were fitted simultaneously to eq 5. The mode of inhibition shifted from metal-competitive to -noncompetitive with pKa = 8.7 ± 0.1, with Ki,comp = 3.7 ± 0.6 μM and Ki,noncomp = 6.6 ± 0.9 μM. Varying [S]/KM ratios and varying proportions of competitive versus noncompetitive inhibition at each pH lead to different apparent Ki values; however, eq 5 takes these into account to yield the true Ki,comp and Ki,noncomp values.

fragment-based inhibitors were used to probe O4’s effect on inhibition (Figure 6a,b). The first pair, pyruvate oxime + Gro3P, lacked the equivalent of DAHP oxime’s C4−O4 group. Gro3P binds in the E4P binding site in a substrate-like orientation.9 The second pair, glyoxylate oxime + Ero4P, lacked C3 but contained the equivalent of the C4−O4 group in the Ero4P molecule. None of the individual fragments inhibited DAHPSH6; inhibition was observed only with the fragment pairs (Figure 6). The modes of inhibition for the oxime-containing fragments were determined using Eadie−Hofstee plots by varying the concentrations of Mn2+ and the oxime-containing fragment while holding the phosphate-containing fragment, Gro3P or Ero4P, constant. The parallel lines for variable pyruvate oxime demonstrated metal-noncompetitive inhibition (Figure 6c), while variable glyoxylate oxime gave lines that converged toward the y-axis, demonstrating metal-competitive inhibition (Figure 6d). The apparent Ki values for the oxime-containing fragments were determined. Because there is no guarantee that the enzyme E

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 4. (a) DAHPS active site residues. The DAHPS·DAHP oxime (PDB ID: 5CKS)18 structure is shown in color. Mn2+, PEP, SO42−, and water (gray) are from DAHPS·Mn2+·PEP·SO42− (PDB ID: 1N8F).30 The crystallographic water in 1N8F occupies the same space as DAHP oxime’s O4 in 5CKS. The side chain positions in 1N8F are similar to 5CKS (see the Supporting Information, Figure S1). The Mn2+-binding site is unoccupied in 5CKS. Two crystallographic water molecules in 5CKS occupy the same position as PEP’s phosphate oxygens in 1N8F. (b) Distances between active site residues and DAHP oxime O4 in 5CKS (teal) or water in 1N8F (gray) (Table S1). Distances in parentheses are between atoms in 5CKS and Mn2+ in 1N8F. (c) Spacefilling model of the interaction between DAHP oxime O4, C61 Sγ, and D326 Oδ2.

Figure 5. DAHP oxime binding to mutant DAHPSs. (a) ITC titration of D326A·Mn2+ with DAHP oxime. (top) Raw ITC trace, (bottom) integrated heat fitted to a one-site binding model, Kd = 14 ± 5 μM. The thin black line is the residual (observed − fitted), offset by −6 μJ. Eadie−Hofstee plots for DAHP oxime inhibition of (b) C61A and (c) H268A.

was saturated with Gro3P or Ero4P, the apparent Ki values were only estimates of the oxime fragments’ affinity. For pyruvate oxime, Ki,noncomp(apparent) = 140 ± 20 μM (eq 7, Figure 6e), while, for glyoxylate oxime, Ki,comp(apparent) = 310 ± 30 μM (eq 6, Figure 6f).

binding with pKa = 8.7 ± 0.1 (Figure 3, Table 1). While there was no direct evidence for the identity of the group responsible for this change, C61 was a strong candidate. Its side chain was required for metal-competitive binding, and the pH dependence of KM,Mn, ascribed to C61’s ionization, had a similar pKa value, 8.2. Mode of Inhibition. Both C61 and D326 were required for metal-competitive inhibition, as the C61A and D326A mutants both exhibited metal-noncompetitive inhibition. Both side chains are located close to DAHP oxime’s O4 atom (Figure 4, Figure S1). The fragment-based inhibitor pairs confirmed that DAHP oxime O4 was essential for metal-competitive binding. The fragment pair lacking O4 (pyruvate oxime + Gro3P) displayed metal-noncompetitive inhibition, while the fragment

■

DISCUSSION pH Dependence of Catalysis and Inhibition. Previously, activity assays, ITC titrations, and crystal structures showed that DAHP oxime and Mn2+ binding were mutually exclusive (i.e., competitive) at pH 7 even though they did not occupy the same physical space. DAHP oxime binding to the DAHPSH6·Mn2+ complex was undetectable by ITC.18 However, in the present study, binding was readily detected at pH 9, making the mode of inhibition metal-noncompetitive at high pH (Figure 2). Equation 5 fitted the initial velocities well, modeling the pHdependent change from metal-competitive to -noncompetitive F

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 6. Fragment-based inhibitor pairs. The fragment pairs (a) pyruvate oxime + Gro3P and (b) glyoxylate oxime + Ero4P. (c, d) Eadie−Hofstee plots to determine the mode of inhibition for the oxime-containing fragment. (e) Pyruvate oxime in the presence of 4 mM Gro3P. (f) Glyoxylate oxime in the presence of 8 mM Ero4P. Relative rates were v0,i/v0, where v0,i was the initial velocity at oxime concentration i and v0 was the control reaction.

with Mn2+, D326 Oδ2, and C61 Sγ (Table S1), meaning that these interactions, in themselves, are favorable. The metal-binding site in 5CKS is “looser” than 1N8F, which is not surprising given the lack of Mn2+, but it does not provide an obvious explanation for the source of competition. The metal-binding residues in the DAHPS·DAHP oxime structure moved away from the Mn2+ site (Figure S1, Table S1). For example, the distance between the axial ligands, C61 and H268, is 4.86 Å with Mn2+ present (1N8F) but 5.65 Å in its absence (5CKS). The D326 and C61 side chains make closer contacts with the crystallographic water in 1N8F than with O4 in 5CKS (Table S1). The lack of obvious, direct structural conflicts between DAHP oxime and Mn2+ again implicates remote interactions. There is further evidence for the importance of remote interactions from the fact that the C61A mutation caused a larger decrease in E4P binding, 440-fold (3.6 kcal mol−1), than could be explained by a direct C61 Sγ···E4P O1 contact.22 The analogous interaction in 1N8F is C61 Sγ···OH2, with a distance of 3.16 Å. This is less than the sum of the S···O van der Waals’ radii, 3.32 Å, and shorter than a normal SH···O hydrogen bond, ≥3.4 Å.31,32 It is consistent with a chalcogen bond, in which the oxygen’s lone pair electrons interact with the C−S bond’s σ*orbital.33,34 Chalcogen bonds and SH···O hydrogen bonds both have relatively low energies, so the putative C61 Sγ···E4P O1 interaction, on its own, could not account for the C61A mutation’s 3.6 kcal mol−1 effect on E4P binding. This implicates

pair that contained it (glyoxylate oxime + Ero4P) was metal competitive. Glyoxylate oxime binding was metal-competitive even without a covalent connection between O4 and the oxime functional group. This implicates some conformational or dynamic change in the enzyme, triggered by the O4/Mn2+/ C61/D326 interaction, controlling the metal/O4 interaction. While it is possible that one of the fragments could bind to DAHPSH6 differently from the corresponding portions of DAHP oxime, and thereby obscure O4’s effect on binding, the consistency between the fragment pair and mutant results strongly argues that the fragments bound similarly to DAHP oxime, and that O4 is a key determinant of the mode of inhibition. Just as removing the side chains that interacted with O4 removed competition with Mn2+, so did removing O4 itself. Structural Origin of Competition. Comparing the structures of DAHPS·DAHP oxime (5CKS)18 and DAHPS· Mn2+·PEP·SO42− (1N8F)30 revealed that the competition between DAHP oxime and Mn2+ must originate from indirect or remote interactions, since a crystallographic water occupies the same position in 1N8F as DAHP oxime O4 does in 5CKS (Figure 4, Figure S1). This is also the probable E4P O1 binding site, as DAHP oxime’s O4 atom is derived from E4P O1. This water molecule does not contain any functional groups to compensate for unfavorable interactions, so its binding must be energetically favorable. It makes contact with all of the atoms implicated in metal-competitive DAHP oxime binding, namely, G

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

containing inhibitors to be missed if, as is commonly done, metal ion concentrations far above KM,metal are used.

other, remote, interactions as contributing to E4P binding and, by analogy, DAHP oxime binding. Contributions from indirect active site interactions is further supported by two further observations. First, competition between DAHP oxime and Mn2+ disappeared at high pH (Figure 3). C61 deprotonation would lead to a higher affinity for Mn2+ and presumably a tightening of the C61 Sγ···Mn2+ distance. This would be consistent with the conformational changes in metal binding in DAHPS with pKa ∼ 8.5 observed by electron paramagnetic resonance spectroscopy.35 Second, glyoxylate oxime binding was metal-competitive even though there is no covalent connection between the inhibitor fragments (glyoxylate oxime and Ero4P O4) nor direct contact between glyoxylate oxime and Mn2+. In order for competitive inhibition to be observed, one of two conditions must be satisfied: (i) the inhibitor and substrate bind to the same enzyme form (mutually exclusively), or (ii) the inhibitor binds first and forms contacts with the same residues as the substrate, disrupting its binding.36 It is not clear which applies in this case. DAHP oxime shares some contacts with Mn2+, including C61 and D326, but the interactions responsible for competitive binding are more indirect, i.e., allosteric. Competitive allosteric inhibition is known, if somewhat uncommon,37,38 though one example is already known with DAHPS, as feedback inhibition of some DAHPSs by phenylalanine appears to be competitive even though the phenylalanine binding site is >20 Å from the active site.11,39 Though the structural explanation for competition is not obvious, it is possible to rationalize why it happens. An E4P O1··· Mn2+ contact is likely essential for catalysis, as it is proposed to polarize the aldehyde functionality toward nucleophilic attack from PEP C3 to form the THI.8,9 However, because product release is rate limiting under common reaction conditions,19 having the DAHP O4···Mn2+ interaction become unfavorable in the enzyme−product complex could accelerate product release. The lack of a structural explanation for DAHP oxime/metal competition implicates other factors, such as changes in protein dynamics or intersubunit interactions. DAHP oxime inhibition already shows a number of unexpected features, including large changes in protein dynamics in spite of minimal structural changes, strong negative cooperativity between inhibitor binding to different subunits, yet strong positive cooperativity between inhibitor binding and substrate binding to uninhibited subunits.18 Metal-competitive DAHP oxime binding is another manifestation of the complex structural and dynamic interactions between DAHPS and its ligands, which has clear implications for inhibitor design. Significance for Inhibitor Design. Regardless of its origin, competition between DAHP oxime and Mn2+ will be an important consideration in inhibitor design, particularly since most potent DAHPS inhibitors contain an O4 group. Bacterial cells contain μM to mM concentrations of several divalent transition metal ions that activate DAHPS.23 This will decrease the in vivo efficiency of any inhibitor that must compete with them. In addition, some metals dissociate extremely slowly from DAHPS, requiring up to 48 h incubations with EDTA to remove.18 Slow-dissociating metals would make metal-competitive inhibitors ineffective. Though removing O4 would entail the loss of hydrogen bonding interactions with K97 and D326, it would be offset by removing the need to compete with intracellular metal ions. Metal competition could also affect inhibitor screening assays and cause otherwise effective O4-

■

CONCLUSIONS Competitive binding between DAHP oxime and the essential metal activator presents a challenge for inhibitor design given bacteria’s sometimes high intracellular transition metal ion concentrations and slow metal ion dissociation from DAHPS. We showed that competition arose due to a four-way interaction between Mn2+, DAHP oxime’s O4 hydroxyl group, and the C61 and D326 side chains. Though the cause of this competition remains elusive, the near ubiquity of the O4 hydroxyl group in DAHP-based inhibitors raises concerns about their efficacy in the cell. Removing the O4 hydroxyl group would likely sacrifice some binding affinity but would remove the problem of metal ion competition, increasing the likelihood of being able to inhibit DAHPS inside bacterial cells.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.8b00837. Derivation of the Eadie−Hofstee plot in the presence of residual activity and comparison of active site structures and distances in the active sites of 1N8F and 5CKS (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Paul J. Berti: 0000-0003-3409-7907 Funding

This work was supported by Canadian Institutes of Health Research Operating Grant MOP-64422 (P.J.B.). Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS We thank Dr. José Carlos Bozelli Jr. and Prof. Richard Epand for help with ITC experiments. ABBREVIATIONS DAHP, deoxy-D-arabinoheptulosonate-7-phosphate; DAHPS, DAHP synthase; DAHPSH6, N-terminally His6-tagged E. coli DAHPS(Phe); E4P, erythrose 4-phosphate; Ero4P, erythritol 4phosphate; Gro3P, glycerol 3-phosphate; K-CAPS, potassium N-cyclohexyl-3-aminopropanesulfonate; K-CHES, potassium N-cyclohexyl-2-aminoethanesulfonate; K-HEPES, potassium 4-(2-hydroxyethyl)-1-piperazineethanesulfonate; IPTG, isopropyl β-D-1-thiogalactopyranoside; ITC, isothermal titration calorimetry; LFER, linear free energy relationship; PEP, phosphoenolpyruvate; Pi, inorganic phosphate; TCEP, tris (2carboxyethyl) phosphine; THI, tetrahedral intermediate

■

REFERENCES

(1) Bravo, I. G., Garcia-Vallve, S., Romeu, A., and Reglero, A. (2004) Prokaryotic origin of cytidylyltransferases and α-ketoacid synthases. Trends Microbiol. 12, 120−128. (2) Reichau, S., Jiao, W., Walker, S. R., Hutton, R. D., Baker, E. N., and Parker, E. J. (2011) Potent inhibitors of a shikimate pathway enzyme

H

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry from Mycobacterium tuberculosis: combining mechanism- and modeling-based design. J. Biol. Chem. 286, 16197−16207. (3) Walker, S. R., Jiao, W., and Parker, E. J. (2011) Synthesis and evaluation of dual site inhibitors of 3-deoxy-D-arabino-heptulosonate 7-phosphate synthase. Bioorg. Med. Chem. Lett. 21, 5092−5097. (4) Walker, S. R., Cumming, H., and Parker, E. J. (2009) Substrate and reaction intermediate mimics as inhibitors of 3-deoxy-D-arabinoheptulosonate 7-phosphate synthase. Org. Biomol. Chem. 7, 3031− 3035. (5) Walker, S. R., and Parker, E. J. (2006) Synthesis and evaluation of a mechanism-based inhibitor of a 3-deoxy-D-arabino heptulosonate 7phosphate synthase. Bioorg. Med. Chem. Lett. 16, 2951−2954. (6) Grison, C., Petek, S., Finance, C., and Coutrot, P. (2005) Synthesis and antibacterial activity of mechanism-based inhibitors of KDO8P synthase and DAH7P synthase. Carbohydr. Res. 340, 529−537. (7) Srinivasan, P. R., and Sprinson, D. B. (1959) 2-Keto-3-deoxy-Darabo-heptonic acid 7-phosphate synthetase. J. Biol. Chem. 234, 716− 722. (8) Shumilin, I. A., Bauerle, R., Wu, J., Woodard, R. W., and Kretsinger, R. H. (2004) Crystal structure of the reaction complex of 3deoxy-D-arabino-heptulosonate-7-phosphate synthase from Thermotoga maritima refines the catalytic mechanism and indicates a new mechanism of allosteric regulation. J. Mol. Biol. 341, 455−466. (9) Konig, V., Pfeil, A., Braus, G. H., and Schneider, T. R. (2004) Substrate and metal complexes of 3-deoxy-D-arabino-heptulosonate-7phosphate synthase from Saccharomyces cerevisiae provide new insights into the catalytic mechanism. J. Mol. Biol. 337, 675−690. (10) Wu, J., and Woodard, R. W. (2006) New insights into the evolutionary links relating to the 3-deoxy-D-arabino-heptulosonate 7phosphate synthase subfamilies. J. Biol. Chem. 281, 4042−4048. (11) Stephens, C. M., and Bauerle, R. (1991) Analysis of the metal requirement of 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase from Escherichia coli. J. Biol. Chem. 266, 20810−20817. (12) Ray, J. M., and Bauerle, R. (1991) Purification and properties of tryptophan-sensitive 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase from Escherichia coli. J. Bacteriol. 173, 1894−1901. (13) McCandliss, R. J., Poling, M. D., and Herrmann, K. M. (1978) 3Deoxy-D-arabino-heptulosonate 7-phosphate synthase. Purification and molecular characterization of the phenylalanine-sensitive isoenzyme from Escherichia coli. J. Biol. Chem. 253, 4259−4265. (14) Sobota, J. M., Gu, M. Z., and Imlay, J. A. (2014) Intracellular Hydrogen Peroxide and Superoxide Poison 3-Deoxy-D-Arabinoheptulosonate 7-Phosphate Synthase, the First Committed Enzyme in the Aromatic Biosynthetic Pathway of Escherichia coli. J. Bacteriol. 196, 1980−1991. (15) Staub, M., and Denes, G. (1969) Purification and properties of the 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase (phenylalanine sensitive) of Escherichia coli K12. I. Purification of enzyme and some of its catalytic properties. Biochim. Biophys. Acta 178, 588−598. (16) Ganson, R. J., and Jensen, R. A. (1988) The essential role of cobalt in the inhibition of the cytosolic isozyme of 3-deoxy-D-arabinoheptulosonate-7-phosphate synthase from Nicotiana silvestris by glyphosate. Arch. Biochem. Biophys. 260, 85−93. (17) Baasov, T., and Knowles, J. R. (1989) Is the first enzyme of the shikimate pathway, 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase (tyrosine sensitive), a copper metalloenzyme? J. Bacteriol. 171, 6155−6160. (18) Balachandran, N., Heimhalt, M., Liuni, P., To, F., Wilson, D. J., Junop, M. S., and Berti, P. J. (2016) Potent Inhibition of 3-Deoxy-Darabinoheptulosonate-7-phosphate (DAHP) Synthase by DAHP Oxime, a Phosphate Group Mimic. Biochemistry 55, 6617−6629. (19) Furdui, C., Zhou, L., Woodard, R. W., and Anderson, K. S. (2004) Insights into the mechanism of 3-deoxy-D-arabino-heptulosonate 7-phosphate synthase (Phe) from Escherichia coli using a transient kinetic analysis. J. Biol. Chem. 279, 45618−45625. (20) Schoner, R., and Herrmann, K. M. (1976) 3-Deoxy-D-arabinoheptulosonate 7-phosphate synthase. Purification, properties, and kinetics of the tyrosine-sensitive isoenzyme from Escherichia coli. J. Biol. Chem. 251, 5440−5447.

(21) DeLeo, A. B., Dayan, J., and Sprinson, D. B. (1973) Purification and kinetics of tyrosine-sensitive 3-deoxy-D-arabino-heptulosonic acid 7-phosphate synthetase from Salmonella. J. Biol. Chem. 248, 2344− 2353. (22) Balachandran, N., To, F., and Berti, P. J. (2017) Linear Free Energy Relationship Analysis of Transition State Mimicry by 3-DeoxyD-arabino-heptulosonate-7-phosphate (DAHP) Oxime, a DAHP Synthase Inhibitor and Phosphate Mimic. Biochemistry 56, 592−601. (23) Finney, L. A., and O’Halloran, T. V. (2003) Transition metal speciation in the cell: insights from the chemistry of metal ion receptors. Science 300, 931−936. (24) Kitagawa, M., Ara, T., Arifuzzaman, M., Ioka-Nakamichi, T., Inamoto, E., Toyonaga, H., and Mori, H. (2006) Complete set of ORF clones of Escherichia coli ASKA library (a complete set of E. coli K-12 ORF archive): unique resources for biological research. DNA Res. 12, 291−299. (25) Sieben, A. S., Perlin, A. S., and Simpson, F. J. (1966) An improved preparative method for D-erythrose 4-phosphate. Can. J. Biochem. 44, 663−669. (26) Lanzetta, P. A., Alvarez, L. J., Reinach, P. S., and Candia, O. A. (1979) An improved assay for nanomole amounts of inorganic phosphate. Anal. Biochem. 100, 95−97. (27) Berti, P. J., and Chindemi, P. (2009) Catalytic Residues and an Electrostatic Sandwich That Promote Enolpyruvyl Shikimate 3Phosphate Synthase (AroA) Catalysis. Biochemistry 48, 3699−3707. (28) Segel, I. H. (1975) Enzyme kinetics: Behaviour and analysis of rapid equilibrium and steady-state enzyme systems, John Wiley and Sons, New York. (29) Wagner, T., Shumilin, I. A., Bauerle, R., and Kretsinger, R. H. (2000) Structure of 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase from Escherichia coli: comparison of the Mn2+ *2phosphoglycolate and the Pb2+*2-phosphoenolpyruvate complexes and implications for catalysis. J. Mol. Biol. 301, 389−399. (30) Shumilin, I. A., Bauerle, R., and Kretsinger, R. H. (2003) The high-resolution structure of 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase reveals a twist in the plane of bound phosphoenolpyruvate. Biochemistry 42, 3766−3776. (31) Zhou, P., Tian, F. F., Lv, F. L., and Shang, Z. C. (2009) Geometric characteristics of hydrogen bonds involving sulfur atoms in proteins. Proteins: Struct., Funct., Genet. 76, 151−163. (32) Gregoret, L. M., Rader, S. D., Fletterick, R. J., and Cohen, F. E. (1991) Hydrogen-Bonds Involving Sulfur-Atoms in Proteins. Proteins: Struct., Funct., Genet. 9, 99−107. (33) Beno, B. R., Yeung, K. S., Bartberger, M. D., Pennington, L. D., and Meanwell, N. A. (2015) A Survey of the Role of Noncovalent Sulfur Interactions in Drug Design. J. Med. Chem. 58, 4383−4438. (34) Iwaoka, M., and Isozumi, N. (2012) Hypervalent Nonbonded Interactions of a Divalent Sulfur Atom. Implications in Protein Architecture and the Functions. Molecules 17, 7266−7283. (35) Jordan, P. A., Bohle, D. S., Ramilo, C. A., and Evans, J. N. (2001) New insights into the metal center of 3-deoxy-D-arabino-heptulosonate 7-phosphate synthase. Biochemistry 40, 8387−8396. (36) Cleland, W. W., and Cook, P. F. (2007) Enyme Kinetics and Mechanism, Garland Science, London. (37) Bindslev, N. (2013) Allosteric transition: a comparison of two models. BMC Pharmacol. Toxicol. 14, 4. (38) Alphey, M. S., Pirrie, L., Torrie, L. S., Boulkeroua, W. A., Gardiner, M., Sarkar, A., Maringer, M., Oehlmann, W., Brenk, R., Scherman, M. S., McNeil, M., Rejzek, M., Field, R. A., Singh, M., Gray, D., Westwood, N. J., and Naismith, J. H. (2013) Allosteric Competitive Inhibitors of the Glucose-1-phosphate Thymidylyltransferase (RmlA) from Pseudomonas aeruginosa. ACS Chem. Biol. 8, 387−396. (39) Webby, C. J., Jaio, W., Hutton, R. D., Blackmore, N. J., Baker, H. M., Baker, E. N., Jameson, G. B., and Parker, E. J. (2010) Synergistic allostery: A sophisticated regulatory network for the control of aromatic amino acid biosynthesis in mycobacterium tuberculosis. J. Biol. Chem. 285, 30567−30576.

I

DOI: 10.1021/acs.biochem.8b00837 Biochemistry XXXX, XXX, XXX−XXX