Deciphering the mechanism of glutaredoxin-catalyzed roGFP2 redox sensing reveals a ternary complex with glutathione for protein disulfide reduction

Abstract

Glutaredoxins catalyze the reduction of disulfides and are key players in redox metabolism and regulation. While important insights were gained regarding the reduction of glutathione disulfide substrates, the mechanism of non-glutathione disulfide reduction remains highly debated. Here we determined the rate constants for the individual redox reactions between PfGrx, a model glutaredoxin from Plasmodium falciparum, and redox-sensitive green fluorescent protein 2 (roGFP2), a model substrate and versatile tool for intracellular redox measurements. We show that the PfGrx-catalyzed oxidation of roGFP2 occurs via a monothiol mechanism and is up to three orders of magnitude faster when roGFP2 and PfGrx are fused. The oxidation kinetics of roGFP2-PfGrx fusion constructs reflect at physiological GSSG concentrations the glutathionylation kinetics of the glutaredoxin moiety, thus allowing intracellular structure-function analysis. Reduction of the roGFP2 disulfide occurs via a monothiol mechanism and involves a ternary complex with GSH and PfGrx. Our study provides the mechanistic basis for understanding roGFP2 redox sensing and challenges previous mechanisms for protein disulfide reduction.

Fig. 1. Overview of the substrates and catalytic mechanisms of class I glutaredoxins.

a Glutaredoxins reduce glutathione- and non-glutathione disulfide substrates using one or two molecules of GSH (shown at the top and bottom, respectively). b Glutaredoxins use a monothiol ping-pong mechanism for the reduction of glutathione disulfide substrates. The glutathione moieties of GSSR and GSH are not identical and have to interact with different protein areas of the reduced and glutathionylated enzyme. c The reduction of selected non-glutathione disulfide substrates via a dithiol mechanism requires a second active-site cysteinyl residue. d NMR-structures of the mixed disulfide between EcGrx1 and a peptide of EcRNR, EcGrx1 disulfide, glutathionylated EcGrx1, and reduced EcGrx1 in accordance with the dithiol mechanism (depicted counterclockwise). The second cysteinyl residue was removed to stabilize the mixed disulfide species. Selected residues that were previously shown to interact with glutathione or to affect catalysis in related enzymes are highlighted. e Three alternative monothiol mechanisms for the reduction of non-glutathione disulfide substrates have been suggested (see text for details). All shown reactions are reversible.

Fig. 2. Spectra and oxidation kinetics for reduced roGFP2^WT, roGFP2^C151S, and roGFP2^C208S.

a Fluorescence spectra of 1 µM roGFP2^WT(SH)₂ and roGFP2^WT(S₂) are shown on the left. The spectrum for roGFP2^WT(S₂) was recorded 2 min after the addition of 20 µM PfGrx^C32S/C88S(SSG). The fluorescence spectra of 1 µM roGFP2^C151S(SH) and roGFP2^C208S(SH) are shown in the middle and on the right, respectively. Both monothiol roGFP2 variants were incubated with 10 mM GSSG to obtain the spectra of roGFP2^C151S(SSG) and roGFP2^C208S(SSG), which were recorded after the removal of GSSG and GSH. b Representative stopped-flow oxidation kinetics for the reaction between 1 µM reduced wild-type or mutant roGFP2 with 20 µM PfGrx^C32S/C88S(SSG). c, d Representative stopped-flow oxidation kinetics at 484 nm for the reaction between 1 µM reduced wild-type or mutant roGFP2 with the indicated concentrations of either PfGrx^C32S/C88S(SSG) or PfGrx^C88S(SSG).

Fig. 3. Rate constants for the oxidation of wild-type and monothiol roGFP2.

Secondary plots for the reaction kinetics of 1 µM reduced wild-type or monothiol roGFP2 with (a) PfGrx^C32S/C88S(SSG), (b) PfGrx^C88S(SSG), or (c) PfGrx^C88S(S₂). d Overview of the second order rate constants k_ox for the oxidation of the indicated roGFP2 variants by the indicated PfGrx mutants. The rate constants are also listed in Table 1. e Data interpretation of the roGFP2 oxidation kinetics. The preferred oxidation pathway is highlighted with thick arrows. All data sets were generated from at least two independent biological replicates.

Fig. 4. Rate constants for the oxidation of roGFP2^WT-PfGrx fusion constructs.

a Representative biphasic stopped-flow oxidation kinetics at 484 nm for the reaction between 1 µM reduced roGFP2^WT-PfGrx^C32S/C88S and 20 µM GSSG. Secondary plots for the k_obs values from the first and second phase of the reaction between GSSG and 1 µM reduced roGFP2^WT-PfGrx fusion construct for (b) PfGrx^WT, (c) PfGrx^C88S, or (d) PfGrx^C32S/C88S. e Overview of the second order rate constants k₁ for the first phase (left) and the GSSG-independent maximum rate constants k₂ for the second phase (right). The rate constants are also listed in Table 1. f Data interpretation of the oxidation kinetics for the three different roGFP2 fusion constructs. The arrows in orange indicate the altered redox environment for roGFP2(SH)₂ due to the glutathionylation of the PfGrx moiety. All data sets were generated from at least two independent biological replicates.

Fig. 5. Rate constants for the deglutathionylation of monothiol roGFP2 variants.

Representative secondary plots for the k_obs values from the reaction between (a) 1 µM roGFP2^C151S(SSG) or (b) 1 µM roGFP2^C208S(SSG) and reduced PfGrx^C88S (top) or monothiol PfGrx^C32S/C88S (bottom) in the absence (left) or presence of GSH (right). c Overview of the second order rate constants k_red for the PfGrx-dependent reduction of the indicated roGFP2(SSG) mutants at four different GSH concentrations. The rate constants are also listed in Table 1 and Supplementary Table S1. d Data interpretation of the reduction kinetics for both monothiol roGFP2(SSG) variants and both PfGrx mutants. The faster reduction pathways are highlighted with thick arrows. All data sets were generated from one or two independent biological replicates.

Fig. 6. Rate constants for the reduction of roGFP2^WT(S₂).

a Representative biphasic stopped-flow reduction kinetics at 484 nm for the reaction between 1 µM roGFP2^WT(S₂) and reduced dithiol PfGrx^C88S (left) or monothiol PfGrx^C32S/C88S (right) in the presence of 5 mM GSH. b, c Representative secondary plots of the k_obs values for the first phase with reduced PfGrx^C88S or PfGrx^C32S/C88S at 1 mM (left) or 5 mM GSH (right). d Overview of the apparent second-order rate constants k_red^app for the first phase of the PfGrx-dependent reduction of roGFP2(S₂) at different GSH concentrations (left) or with 1 mM GSH in the presence or absence of the competitive inhibitor GSMe (right). No activity was detected for 0 mM GSH. The rate constants are also listed in Table 1 and Supplementary Table S1. e Secondary plots for the first phase of the reduction kinetics of roGFP2(S₂) at 5 mM GSH using either a D90A (left) or K26A (right) mutant of reduced monothiol PfGrx^C32S/C88S. f Tertiary plots of the GSH-dependent k_red^app values for reduced PfGrx^C88S (left) or PfGrx^C32S/C88S (right). True k_red values were estimated by linear regression and are listed in Table 1. All data sets were generated from at least two independent biological replicates.

Fig. 7. Simulations of the stopped-flow kinetic data for the reduction of roGFP2^WT(S₂).

Simulation of the biphasic kinetic data for the reaction between 1 μM roGFP2^WT(S₂) and 1, 2.5, or 5 mM GSH in the presence of 1–40 µM reduced PfGrx^C88S. Best fits (dashed lines) were obtained for a reaction within a ternary complex regardless whether (a) roGFP2(S₂) or (b) GSH was first bound to PfGrx. Rounded rate constants from Table 1 in black were chosen as input parameters to calculate the missing parameters in blue. The relative fluorescence intensity of each roGFP2 redox species was taken from Fig. 2.

Fig. 8. Potential effects of the transition state geometries and substrate sizes on the glutaredoxin-catalyzed reduction of non-glutathione disulfide substrates.

a Cleland scheme of the deduced monothiol reaction mechanism for the PfGrx-catalyzed redox reactions of roGFP2. The reaction with roGFP2(S₂) as a model non-glutathione disulfide substrate begins with a rate-limiting random bi-uni mechanism and is shown from left to right. The subsequent deglutathionylation of the roGFP2(SSG) intermediate follows a ping-pong mechanism. The reverse reaction with GSSG (or GSSR) and roGFP2(SH)₂ as model protein with surface-exposed cysteinyl residues is shown from right to left and will probably deviate by a nonenzymatic shortcut yielding roGFP2(S₂) from roGFP2(SSG). See text for details. b Schematic representations of the three predicted transition states for the reactions from panel a. c Potential effects of the size and reaction geometry on the formation and orientation of the first transition state from (a). d Schematic representations of the four predicted transition states for the suggested reduction of HEDS according to monothiol mechanism (i). e Schematic representations for the putative reduction of Grx(SSR) intermediates according to alternative monothiol mechanism (i) are shown on the left and in the middle. If the Grx(SSR) species cannot be attacked by GSH, a second or moderately conserved third cysteinyl residue in glutaredoxins could prevent the accumulation of trapped Grx(SSR) species according to the cysteine resolving model shown on the right.