S-nitrosylation of Hsp90 promotes the inhibition of its ATPase and endothelial nitric oxide synthase regulatory activities

Abstract

Nitric oxide is implicated in a variety of signaling pathways in different systems, notably in endothelial cells. Some of its effects can be exerted through covalent modifications of proteins and, among these modifications, increasing attention is being paid to S-nitrosylation as a signaling mechanism. In this work, we show by a variety of methods (ozone chemiluminescence, biotin switch, and mass spectrometry) that the molecular chaperone Hsp90 is a target of S-nitrosylation and identify a susceptible cysteine residue in the region of the C-terminal domain that interacts with endothelial nitric oxide synthase (eNOS). We also show that the modification occurs in endothelial cells when they are treated with S-nitroso-l-cysteine and when they are exposed to eNOS activators. Hsp90 ATPase activity and its positive effect on eNOS activity are both inhibited by S-nitrosylation. Together, these data suggest that S-nitrosylation may functionally regulate the general activities of Hsp90 and provide a feedback mechanism for limiting eNOS activation.

Fig. 1.

Detection of S-nitrosylation by the biotin switch technique and by specific reduction and ozone chemiluminescence. (A) Recombinant human Hsp90α (1.3 μM) was left untreated (-) or treated with 1 mM diethylamine-NONOate (DEA/NO), 2.7 μM NaNO₂ in 0.5 M HCl (NO₂^-/H⁺), or 1 mM GSNO. Hsp90α preparations were then subjected to the biotin switch assay, and biotinylated protein was detected by Western blot with streptavidin. The image shown is representative of three independent experiments. (B and C) Recombinant human Hsp90α was reduced with Cu⁺/cysteine, and NO release detected in a nitric oxide analyzer apparatus. Figures shown are representative of at least three independent experiments. (B) Protein (5.3 μM) or buffer was treated with 1 mM GSNO, and protein and low molecular-mass (LMM) fractions were separated in a Sephadex G-25 column before analysis. 1: Hsp90 plus GSNO, protein fraction; 2: Hsp90 plus GSNO, LMM fraction; 3: Hsp90, protein fraction; 4: buffer plus GSNO, protein fraction; 5: buffer plus GSNO, LMM fraction. (C) Protein (0.8 μM) was treated with 1.6 μM NaNO₂ in 0.5 M HCl or HCl only for 20 min and adjusted to pH 7.4 with NaOH and PBS before analysis. 1: Buffer plus NaNO₂; 2: protein without NaNO₂; 3: protein plus NaNO₂.

Fig. 2.

Detection and identification of S-nitrosylation by mass spectrometry. (A) Recombinant human Hsp90α was either treated with NaNO₂ (1:2 molar ratio) in 0.5 M HCl and adjusted to pH 7.4 with NaOH and PBS or with 1 mM GSNO in PBS. Samples were then trypsin digested and analyzed by LC-MS by using a linear ion trap detector, as described in Materials and Methods. Shown is the fragmentation spectrum of the S-nitrosylated peptide corresponding to residues 591–611 of Hsp90α, together with the assignation of the fragmentation series to the sequence of the modified peptide. *, tentative S-nitrosylated Cys. (B) Schematic of Hsp90α domains (in boxes) and the localization of the cysteine residues (dots) and the S-nitrosylated peptide detected by MS. Functions assigned to the domains are written above, and regions described to interact with Akt and eNOS are shown below. Numbers correspond to the human Hsp90α sequence and, in the case of the domains, they denote the boundaries of the crystallographic structures. (C) Ribbon representation of the C-domain structure model of human Hsp90α (residues 568 to 693 of swiss-prot entry P07900), based on the structure of the C-terminal domain of E. coli htpG (PDB ID code 1fs8D). The side chains of cysteines 596 and 597, arginine 590, and glutamic acid 634 are highlighted.

Fig. 3.

Detection of S-nitrosylated Hsp90 in endothelial cells. (A) EA.hy926 cells were untreated (-) or exposed (+) to 1 mM cysteine (Cys) or S-nitrosocysteine (CSNO) for 15 min. Protein extracts were subjected to the biotin switch technique, and the biotinylated proteins were purified with avidin-agarose, followed by elution with 2-mercaptoethanol. Hsp90 was detected in total protein extracts and in eluates by SDS/PAGE and immunoblot. (B) EA.hy926 cells were untreated, treated with Cys or CSNO (1 mM), or stimulated with VEGF (50 ng/ml) with or without prior incubation with L-NAME (0.5 mM, 1 h) or calcium ionophore A23187 (10 mM) for 15 min. Protein extracts were immunoprecipitated with anti S-nitrosocysteine and Hsp90 identified by immunoblot. (C) Extracts from EA.hy926 cells untreated or exposed to Cys or CSNO (1 mM) were immunoprecipitated with anti-S-nitrosocysteine with or without prior specific breakdown of the S-NO bond by Hg²⁺ and Hsp90 identified by immunoblot. Figures shown are representative of at least two independent experiments.

Fig. 4.

Hsp90 ATPase activity is inhibited by S-nitrosylation. Hsp90 was 1 μM throughout. (A) Activities are shown for Hsp90 alone, after S-nitrosylation with 1 mM GSNO (n = 7), or incubation with the Hsp90-specific inhibitor, geldanamycin (GA; 2 μM, n = 4). ATPase activity is expressed as a relative slope to that of Hsp90 alone. Values are means ± SEM; *, P < 0.05 with respect to Hsp90 alone. (B) A representative experiment showing experimental progress curves corresponding to NADH consumption after addition of Hsp90 alone (in Tris buffer), Hsp90 plus a GA vehicle (DMSO), Hsp90 preincubated with GSNO, or Hsp90 plus GA.

Fig. 5.

eNOS activation by Hsp90 is inhibited by S-nitrosylation. Purified eNOS (1 μM) was incubated with or without native or S-nitrosylated Hsp90 (1 μM), and eNOS activity was determined from the conversion of ³H-l-arginine to ³H-l-citrulline. Data are means ± SEM, n = 7 experiments in duplicate; *, P < 0.01 with respect to eNOS alone.