GAPDH mediates nitrosylation of nuclear proteins

Abstract

S-nitrosylation of proteins by nitric oxide is a major mode of signalling in cells. S-nitrosylation can mediate the regulation of a range of proteins, including prominent nuclear proteins, such as HDAC2 (ref. 2) and PARP1 (ref. 3). The high reactivity of the nitric oxide group with protein thiols, but the selective nature of nitrosylation within the cell, implies the existence of targeting mechanisms. Specificity of nitric oxide signalling is often achieved by the binding of nitric oxide synthase (NOS) to target proteins, either directly or through scaffolding proteins such as PSD-95 (ref. 5) and CAPON. As the three principal isoforms of NOS--neuronal NOS (nNOS), endothelial NOS (eNOS) and inducible NOS (iNOS)--are primarily non-nuclear, the mechanisms by which nuclear proteins are selectively nitrosylated have been elusive. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is physiologically nitrosylated at its Cys 150 residue. Nitrosylated GAPDH (SNO-GAPDH) binds to Siah1, which possesses a nuclear localization signal, and is transported to the nucleus. Here, we show that SNO-GAPDH physiologically transnitrosylates nuclear proteins, including the deacetylating enzyme sirtuin-1 (SIRT1), histone deacetylase-2 (HDAC2) and DNA-activated protein kinase (DNA-PK). Our findings reveal a novel mechanism for targeted nitrosylation of nuclear proteins and suggest that protein-protein transfer of nitric oxide groups may be a general mechanism in cellular signal transduction.

Figure 1. SNO-GAPDH interacts with SIRT1 near its nitrosylated Cys150 residue

(a) Endogenous co-immunoprecipitation of SIRT1 and GAPDH in HEK293 cells treated with NO donor. Cells were treated with 200 μM GSH or GSNO for 16 hr prior to lysis. (b) Nitrosylated GAPDH (SNO-GAPDH) binds directly to SIRT1 in vitro. GST or GST-GAPDH was pre-treated with 100 μM GSH or GSNO for 30 min at 37°C. After desalting, recombinant SIRT1 was added and binding assessed by a GSH-agarose pulldown assay. (c) A small peptide corresponding to the region of GAPDH that spans Cys150 (Peptide-C150) blocks the interaction between SNO-GAPDH and SIRT1. The assay was performed as in b. (d) Mutation of Thr152 of GAPDH abolishes binding to SIRT1. Twenty-four hours after transfection with wild-type HA-GAPDH or the indicated point mutants, HEK293 cells were treated with 200 μM GSH or GSNO for 16 hr. Cell lysates were immunoprecipitated with anti-SIRT1 antibody and analyzed by western blotting with anti-HA antibody. HA-GAPDH, HA-tagged GAPDH.

Figure 2. Nuclear SNO-GAPDH mediates nitrosylation of SIRT1 via transnitrosylation

(a) Endogenous SIRT1 is nitrosylated in 293-nNOS cells treated with the calcium ionophore A23187 (5 μM, 2 hr). (b) SIRT1 is nitrosylated in cortical neurons treated with NMDA. Nitrosylation is abolished by pre-treatment with the nNOS inhibitor L-VNIO (100 μM, 2 hr). (c) In vitro transnitrosylation assay. Recombinant SIRT1 was incubated with recombinant GAPDH (wild-type, T152A, or C150S, approx. 0.75 μM) that had been pre-treated with 50 μM GSH or GSNO and desalted to remove excess small molecules. A biotin switch assay was then performed. (d) Overexpression of GAPDH augments SIRT1 nitrosylation in 293-nNOS cells. (e) Mutation of Thr152 or Cys150 abrogates the effect of GAPDH overexpression. (f) Depletion of GAPDH by RNAi in 293-nNOS cells leads to a loss of nitrosylation of endogenous SIRT1. (g) The effect of GAPDH knockdown on SIRT1 nitrosylation is rescued by wild-type GAPDH but not GAPDH-T152A. (h) Siah1^ΔNLS prevents NO-induced nuclear translocation of GAPDH. 293T cells were transfected with the indicated plasmid and treated with or without 500 μM GSNO for 3 hr followed by nuclear fractionation. (i) Siah1^ΔNLS abolishes nitrosylation of endogenous SIRT1 in 293-nNOS cells. FH-SIRT1, Flag-HA-tagged SIRT1. Where indicated, cells were treated with 5 μM A23187 for 2 hr. Results from this figure are quantified in Supplementary Information, Fig. S2.

Figure 3. SNO-GAPDH mediates inhibition of SIRT1 enzymatic activity by nitric oxide

(a) GSNO treatment inhibits SIRT1 catalytic activity in an in vitro histone deacetylation assay. SIRT1 was pre-treated with the indicated concentration of GSH or GSNO and desalted prior to the assay. (b) Pre-incubation with SNO-GAPDH inhibits SIRT1 catalytic activity in vitro. This effect is lost with mutation of GAPDH-T152. The assay was performed as in a. *P<0.05, n=3, mean±s.e.m., one-way ANOVA. (c) Activation of nNOS leads to increased acetylation of PGC1α. Twenty-four hours after transfection with Flag-PGC1α, HEK293 or 293-nNOS cells were treated with A23187 (5 μM) for the indicated times prior to lysis. (d) Depletion of GAPDH in 293-nNOS cells by RNAi leads to decreased acetylation of PGC1α. Cells were treated with A23187 (5 μM) for 6 hr prior to lysis. *P<0.01, n=3, mean ± s.e.m., student’s t-test. (e) GAPDH regulates PGC1α/HNF4α transcriptional activity via nitrosylation of SIRT1. Depletion of GAPDH in 293-nNOS cells transfected with plasmids encoding HNF4α and an HNF4α binding site luciferase leads to increased luciferase activity. This effect is rescued by wild-type GAPDH but not GAPDH-T152A. Cells were treated with 5 μM A23187 for 3 hr. *P<0.05, n=3, mean ± s.e.m., student’s t-test. Flag-PGC1α, Flag-tagged PGC1α.

Figure 4. Identification of HDAC2 and DNA-PK as nuclear targets of SNO-GAPDH mediated transnitrosylation

(a) Endogenous co-immunoprecipitation of HDAC2 and GAPDH in HEK293 cells treated with NO donor. Cells were treated with 200 μM GSH or GSNO for 16 hr prior to lysis. (b) and (c) GAPDH knockdown (b) and Siah1^ΔNLS expression (c) lead to a loss of HDAC2 nitrosylation in 293- nNOS cells. (d) and (e) Similar results were obtained for DNA-PK. *P<0.05, n=3, mean±s.e.m., student’s t-test.