Communication between thiamin cofactors in the Escherichia coli pyruvate dehydrogenase complex E1 component active centers: evidence for a "direct pathway" between the 4'-aminopyrimidine N1' atoms

Abstract

Kinetic, spectroscopic, and structural analysis tested the hypothesis that a chain of residues connecting the 4'-aminopyrimidine N1' atoms of thiamin diphosphates (ThDPs) in the two active centers of the Escherichia coli pyruvate dehydrogenase complex E1 component provides a signal transduction pathway. Substitution of the three acidic residues (Glu(571), Glu(235), and Glu(237)) and Arg(606) resulted in impaired binding of the second ThDP, once the first active center was filled, suggesting a pathway for communication between the two ThDPs. 1) Steady-state kinetic and fluorescence quenching studies revealed that upon E571A, E235A, E237A, and R606A substitutions, ThDP binding in the second active center was affected. 2) Analysis of the kinetics of thiazolium C2 hydrogen/deuterium exchange of enzyme-bound ThDP suggests half-of-the-sites reactivity for the E1 component, with fast (activated site) and slow exchanging sites (dormant site). The E235A and E571A variants gave no evidence for the slow exchanging site, indicating that only one of two active sites is filled with ThDP. 3) Titration of the E235A and E237A variants with methyl acetylphosphonate monitored by circular dichroism suggested that only half of the active sites were filled with a covalent predecarboxylation intermediate analog. 4) Crystal structures of E235A and E571A in complex with ThDP revealed the structural basis for the spectroscopic and kinetic observations and showed that either substitution affects cofactor binding, despite the fact that Glu(235) makes no direct contact with the cofactor. The role of the conserved Glu(571) residue in both catalysis and cofactor orientation is revealed by the combined results for the first time.

FIGURE 1.

Shown are the amino acids of E1ec that connect one ThDP cofactor to the other. Residue Glu⁵⁷¹ (E571) is the highly conserved glutamate present in ThDP-dependent enzymes and forms a hydrogen bond directly with the ThDP cofactor.

FIGURE 2.

Steady-state analysis of ThDP binding at 30 °C. The reaction mixture contained the following in 1 ml: 0.10 m Tris-HCl (pH 8.0), 0.2 mm MgCl₂, 2.5 mm NAD⁺, 2.6 mm dithiothreitol, and 2.0 mm sodium pyruvate. The reaction was started by the addition of 0.20 mm CoA and 3 μg of PDHc, resulting from reconstitution of E1ec or its variant with E2ec-E3ec subcomplex at a mass ratio of 1:1:1. The insets show the dependence of the lag phase on ThDP concentration.

SCHEME 1.

Mechanism of pyruvate dehydrogenase.

FIGURE 3.

Dependence of the percentage of fluorescence quenching at 330 nm on the concentration of ThDP. E1ec and its variants (0.034 mg/ml, concentration of active centers = 0.342 μm) in 10 mm KH₂PO₄ (pH 7.0) containing MgCl₂ (5 mm) and pyruvate (1.0 mm) were titrated with ThDP (0.5–300 μm). The values of K_d,ThDP were calculated using an equation for ligand binding at two active centers with the Sigma Plot 10.0 program and are presented in Table 4.

FIGURE 4.

Dependence of the quenching of fluorescence at 330 nm on N1′-methyl-ThDP concentration. Quenching of intrinsic fluorescence of E1ec showing no saturation (top). Quenching of the intrinsic fluorescence of E571A E1ec showing saturation with K_d = 66 μm (bottom). The conditions for both experiments were as follows. E1ec (0.046 mg/ml, concentration of active centers = 0.23 μm) and E571A E1ec (0.084 mg/ml, concentration of active centers = 0.42 μm) in 10 mm KH₂PO₄ (pH 7.0) containing 5 mm MgCl₂ and 2 mm pyruvate were titrated by 1–300 μm N1′-methyl-ThDP.

FIGURE 5.

Hydrogen/deuterium exchange kinetics of C2-H of E1ec-bound ThDP. Top, expansions of selected one-dimensional ¹H NMR spectra of acid quench-isolated ThDP after reaction of 1 volume of E1ec (reconstituted with an equimolar amount of ThDP and 1 mm Mg²⁺) in 20 mm KH₂PO₄ (pH 7.0) with 1 volume of 99.9% D₂O for different reaction times at 30 °C. The singlet signals of C2-H (9.68 ppm) and C6′-H (8.01 ppm) are shown. The latter signal serves as an internal non-exchanging standard for quantification. Bottom, the decay of the ¹H NMR integral intensity of the ThDP C2-H signal relative to the integral of the non-exchanging C6′-H signal was fitted to a double exponential function for exchange of enzyme-bound ThDP (○) as detailed in Ref. . The estimated pseudo-first order rate constants are summarized in Table 5.

FIGURE 6.

Hydrogen/deuterium exchange kinetics of C2-H of ThDP for E1ec variants. Top, E571A; middle, E235A; bottom, R606A. NMR data were fitted to either a double exponential (E235A/E571A) or monoexponential (R606A) function. The estimated pseudo-first order rate constants are summarized in Table 5.

SCHEME 2.

Mechanism of formation of predecarboxylation intermediate and its stable analog.

FIGURE 7.

Near-UV CD spectra of E1ec variants upon titration by methyl acetylphosphonate. The E235A (top) or E237A (bottom) variants (2.0 mg/ml; concentration of active centers = 20 μm) in 20 mm KH₂PO₄ (pH 7.0) containing 2 mm MgCl₂ and 0.50 mm ThDP were titrated with 1–100 μm MAP. The values of K_d,MAP were calculated using a quadratic equation (11). The inset shows that the limiting value of the CD band is reached at a 0.5:1 molar ratio of [MAP]/[E1ec variant active centers].

FIGURE 8.

Rates of the 1′,4′-iminophosphonolactyl ThDP formation for E1ec and its variants determined by time-resolved circular dichroism. E1ec or its variants (5–10 mg/ml) in 20 mm KH₂PO₄ (pH 7.0) containing ThDP (0.50 mm) and MgCl₂ (2.0 mm) was mixed with MAP (2 mm) in the same buffer, and reaction was monitored for varied time intervals (5–200 s). The data were treated as a double exponential as in Equation 1.

FIGURE 9.

Circular dichroism spectra of E571A and E235A variants upon titration by ThTTDP. E571A (top, 2.1 mg/ml, concentration of active centers = 21 μm) and E235A (bottom, 1.3 mg/ml, concentration of active centers = 6.3 μm) in 20 mm KH₂PO₄ (pH 7.0) containing 2 mm MgCl₂ were titrated by 1–90 μm ThTTDP. The insets show the dependence of the circular dichroism at the maximum on concentration of ThTTDP. Data for E235A were treated using the Hill equation (11).

FIGURE 10.

Stereo 2F_o − F_c omit electron density map contoured at 1σ for the dimer interface environment around the catalytic centers (upper images in subunit a and lower images in subunit b) of the E571A-ThDP complex. Coordinates for the thiazolium and pyrimidine rings were used from the E1ec-ThDP structure to create the complete cofactor model in the active centers for illustrative purposes. The electron density map indicates dual disordered positions for the His¹⁴² and Tyr⁵⁹⁸ residues.

FIGURE 11.

2F_o − F_c omit electron density map contoured at 1σ for the ThDP, Glu²³⁵, Glu⁵⁷¹, and Arg⁶⁰⁶ residues on the proposed communication path for E1ec (top) and E235A (concentration of ThDP = 0.5 mm) (bottom). The active center with the disordered His¹⁴² is shown. The electron density is colored red for the ThDP and blue for protein residues.