Dehydroglutathione, a glutathione derivative to introduce non-reversible glutathionylation

Abstract

Protein cysteine is susceptible to diverse oxidations, including disulfide, S-sulfenylation, S-nitrosylation, and S-glutathionylation, that regulate many biological processes in physiology and diseases. Despite evidence supporting distinct biological outcomes of individual cysteine oxoforms, the approach for examining functional effects resulting from a specific cysteine oxoform, such as S-glutathionylation, remains limited. In this report, we devised a dehydroglutathione (dhG)-mediated strategy, named G-PROV, that introduces a non-reducible glutathionylation mimic to the protein with the subsequent delivery of the modified protein to cells to examine the "phenotype" attributed to "glutathionylation". We applied our strategy to fatty acid binding protein 5 (FABP5), demonstrating that dhG induces selective modification at C127 of FABP5, resembling S-glutathionylation. dhG-modified glutathionylation in FABP5 increases its binding affinity to linoleic acid, enhances its translocation to the nucleus for activating PPARβ/δ, and promotes MCF7 cell migration in response to linoleic acid. Our data report a facile chemical tool to introduce a glutathionylation mimic to proteins for functional analysis of protein glutathionylation.

Fig. 1. G-PROV approach for functional study of glutathionylated proteins in cells. The protein of interest (POI) is subjected to a reaction with dehydroglutathione (dhG) in vitro, which induces a non-reducible mimic of glutathionylation (step 1). The dhG-mediated glutathione-modified POI is delivered to the cytoplasm of cells via a fusogenic liposome for functional phenotype analysis (step 2).

Fig. 2. dhG reaction with Cys results in a thioether linkage of glutathione modification. (A) dhG reaction with N-acetylcysteine (NAC), resulting in the Michael reaction adduct. (B) The reaction of a Cys-containing peptide (PEP) with dhG. PEP (0.1 mM) was incubated without (top) or with dhG (1 mM) (bottom) in PBS (pH 8) for 1 h, and was analyzed by HPLC-MS (monitoring absorbance at 214 nm). (C) The reaction of fluorescein-conjugated PEP (FAM-PEP) with dhG in urea-gel electrophoresis (n = 3, biological replicates). FAM-PEP and its conjugation product were monitored by fluorescence. (D) dhG reaction kinetics with FAM-PEP. The intensities of the FAM-PEP bands at different times in urea-gels were plotted for the rate (top). Reactions were assumed to follow pseudo-first-order kinetics. Reaction rates were plotted as a function of dhG concentration (bottom) to determine the second-order rate constant (n = 3, biological replicates). Data show the mean ± SD (D) and are representative of replicate experiments (B) and (C).

Fig. 3. dhG induces glutathione modification on C127 of FABP5. (A) FABP5 structure (PDB: 4LKT) with positions of 6 cysteine residues. FABP5 has a twisted β-barrel structure with two helices (α1 and α2) acting as a lid (left). An enlarged structure around C127 with residues in proximity (middle). The size and depth of the lipid-binding pocket in FABP5 (right). Linoleic acid is shown in a stick model (orange). (B) dhG-modification on FABP5 WT. Increasing amounts of dhG were incubated with purified FABP5 in PBS, which was analyzed by Coomassie stain (CM) and glutathione antibody (GSH) (n = 2, biological replicates). (C) dhG-modification on FABP5 WT and cysteine mutants (n = 3, biological replicates). (D) GSSG-mediated S-glutathionylation of FABP5 WT and cysteine mutants. Purified FABP5 constructs were incubated with GSSG for 1 h (n = 3, biological replicates). (E) MALDI-TOF analysis of FABP5 WT or C127S incubated with dhG or GSSG. FABP5 constructs were incubated with dhG (10 mM) or GSSG (5 mM) for 1 h (n = 3, biological replicates). (F) MS2 spectrum of a dhG-modified peptide in FABP5. FABP5 modified by dhG was digested by CNBr and analyzed by LC-MS/MS, finding a peptide modified by dhG at C127. Data show the mean ± SD (B)–(D) and are representative of replicate experiments (B)–(F). The statistical difference was analyzed by one-way (B) and (C) or two-way (D) ANOVA with Tukey's post hoc test, where *p < 0.03, **p < 0.002, ***p < 0.0002, ****p < 0.0001.

Fig. 4. dhG-modification of FABP5 increases the binding affinity with linoleic acid. (A) and (B) The isothermal titration calorimetry (ITC) to measure FABP5 binding affinity with linoleic acid (LA) upon dhG modification. FABP5 WT (A) and C120S (B) without or with incubation of dhG (10 mM) were purified before the measurement by ITC (n = 3, biological replicates). The differential power (DP) was measured while LA (1 mM) was added to FABP5 (0.1 mM) in Tris–HCl, pH 7.4, over time. (C) The summary of the binding affinity between LA and FABP5 constructs without or with modification by dhG or GSSG. (D) The thermodynamic parameters of FABP5 binding interactions to LA. Data show the mean ± SD (C) and (D) and are representative of replicate experiments (A) and (B). The statistical difference was analyzed by two-way ANOVA with Tukey's post-hoc test (D), where *p < 0.03, **p < 0.002, ***p < 0.0002, ****p < 0.0001.

Fig. 5. FABP5 glutathione modification increases its nuclear translocation, PPARβ/δ activation, and MCF7 cell migration in response to linoleic acid. Fusogenic liposomes alone (FL) or containing FABP5 WT or C127S without or with dhG modification (i.e., WT, C127S, WT-SG, or C127S-SG) were incubated in MCF7 cells for 1 h. (A) Analysis of FABP5 in MCF7 cells after incubating fusogenic liposomes. Lysates were analyzed by western blots (n = 3, biological replicates). (B) Localization of FLAG-FABP5 upon adding linoleic acid (LA). After incubation of LA for 1 h, FABP5 localization (FLAG, green) was analyzed by immunostaining along with DAPI (blue) (n = 10 images out of 2 biological replicates). A scale bar = 10 μm. (C) FABP5 nuclear level upon adding LA. MCF7 cells were treated with none or LA for 12 h. MCF7 cells were lysed, and nuclear extracts were analyzed by western blots (n = 2, biological replicates). (D) PPARβ/δ activation upon adding LA. After incubating LA for 12 h, the nuclear extracts were collected, and the levels of PPARβ/δ bound to the peroxisome proliferator response element (PPRE) were measured by absorbance (n = 4, biological replicates). (D) The in vitro scratch migration assays of MCF7 cells upon incubating LA for 24 h. After incubating fusogenic liposomes containing FABP5 WT, MCF7 cells were incubated without or with LA. The images were taken at 0 and 24 h (n = 5 images out of 3 biological replicates). Yellow colors indicate the area without cells. A scale bar = 0.5 mm. Data show the mean ± SD (A)–(E) and are representative of replicate experiments (A)–(C) and (E). The statistical difference was analyzed by one-way (A) and (B) or two-way (C)–(E) ANOVA and Tukey's post-hoc test, where *p < 0.03, **p < 0.002, ***p < 0.0002, ****p < 0.0001.