Nitro-fatty acids as activators of hSIRT6 deacetylase activity

Abstract

Sirtuin 6, SIRT6, is critical for both glucose and lipid homeostasis and is involved in maintaining genomic stability under conditions of oxidative DNA damage such as those observed in age-related diseases. There is an intense search for modulators of SIRT6 activity, however, not many specific activators have been reported. Long acyl-chain fatty acids have been shown to increase the weak in vitro deacetylase activity of SIRT6 but this effect is modest at best. Herein we report that electrophilic nitro-fatty acids (nitro-oleic acid and nitro-conjugated linoleic acid) potently activate SIRT6. Binding of the nitro-fatty acid to the hydrophobic crevice of the SIRT6 active site exerted a moderate activation (2-fold at 20 μm), similar to that previously reported for non-nitrated fatty acids. However, covalent Michael adduct formation with Cys-18, a residue present at the N terminus of SIRT6 but absent from other isoforms, induced a conformational change that resulted in a much stronger activation (40-fold at 20 μm). Molecular modeling of the resulting Michael adduct suggested stabilization of the co-substrate and acyl-binding loops as a possible additional mechanism of SIRT6 activation by the nitro-fatty acid. Importantly, treatment of cells with nitro-oleic acid promoted H3K9 deacetylation, whereas oleic acid had no effect. Altogether, our results show that nitrated fatty acids can be considered a valuable tool for specific SIRT6 activation, and that SIRT6 should be considered as a molecular target for in vivo actions of these anti-inflammatory nitro-lipids.

Keywords: Michael addition; SIRT6; activator; cysteine covalent modification; docking; enzyme activation; enzyme activator; enzyme kinetics; enzyme mechanism; enzyme structure; fatty acid; fatty acids; histone deacetylase (HDAC); histone deacetylation; nitro-fatty acid; nitro-fatty acids; sirtuin.

Figure 1

Human SIRT6 structure. Molecular representation of hSIRT6 in complex with a myristoylated H3 peptide at Lys-9 (H3K9Myr) and a co-substrate (PDB code 3ZG6). Structural features such as functional domains, A, B, C pockets and relevant residues are indicated. The location of N and C termini (N_T and C_T, respectively) is also shown. Continuation of the peptide chain structure toward C_T is sketched as a guide for helping visualization.

Figure 2

Effect of fatty acids and nitro-fatty acids on in vitro deacetylase activity of hSIRT6.A, chemical structures of 9- and 10-nitro-oleic acids (NO₂-OA), 9- and 12-nitro-conjugated linoleic acids (NO₂-CLA). Experiments were performed using the racemic mixtures 9- and 10-NO₂-OA, and 9- and 12-NO₂-CLA in the presence of 1 μm SIRT6. B, representative runs of deacetylase activity measured using the coupled assay for untreated hSIRT6 (black), hSIRT6 incubated with oleic acid (50 μm, light gray), and hSIRT6 incubated with nitro-oleic acid (20 μm, dark gray). C, fold-change in SIRT6 deacetylase activity after incubation with the indicated concentrations of OA (○), NO₂-OA (●), and NO₂-CLA (■). Error bars represent S.D. of at least three replicates. *, p < 0.1; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001 ANOVA test.

Figure 3

CD of hSIRT6. Far-UV (A) and near-UV (B) CD spectra of hSIRT6 (3 and 10 μm, respectively) after 30 min incubation at 37 °C in the absence (black) or presence of either 5-fold excess of OA (light gray) or NO₂-OA (dark gray). Spectra were normalized for comparison as detailed under “Experimental procedures.”

Figure 4

Consumption of NO₂-CLA by hSIRT6.A, UV-visible spectra of hSIRT6 (solid black line) with a characteristic peak at 280 nm, NO₂-CLA alone (dashed line) with a characteristic peak at 330 nm, and of a mixture of 15 μm hSIRT6 with 10 μm NO₂-CLA. B, change in absorbance at 330 nm with time after mixing hSIRT6 with NO₂-CLA (1.5:1).

Figure 5

Identification of hSIRT6 residues modified by NO₂-OA.A, untreated hSIRT6 (control), and hSIRT6 treated with either NEM or DEPC to block cysteine and histidine residues, respectively, was incubated with the indicated excess of biotinylated-NO₂-OA and adduct formation detected using streptavidin-HRP (upper panel). Ponceau S staining was used as a loading control (lower panel) for densitometric analysis: control (black), NEM-treated SIRT6 (gray), and DEPC-treated SIRT6 (white). One of two consistent experiments is shown. B, MS spectra of untreated-hSIRT6 peptides obtained from tryptic digestion (2 h at 37 °C). The upper panel corresponds to the fragmentation of the doubly charged peptide sequence containing Cys-18 (CGLPEIFDPPEELER; monoisotopic m/z: 872.73) and the lower panel corresponds to triply charged sequence containing Cys-141 (LAELHGNMFVEECAK; monoisotopic m/z: 564.34). C, MS/MS spectra of NO₂-OA–treated SIRT6 (×5) ions with Δm = 327.24. Upper panel corresponds to the doubly charged peptide sequence containing Cys-18 (CGLPEIFDPPEELER; monoisotopic m/z: 1036.21) and lower panel corresponds to the triply charged sequence containing Cys-141 (LAELHGNMFVEECAK; monoisotopic m/z: 673.56). The spectra are representative of three independent experiments. y- and b-ions detected by MS/MS analysis for each tryptic peptide.

Figure 6

Effect of fatty acids and nitro-fatty acids on in vitro deacetylase activity of WT and C18S hSIRT6. Fold-change in SIRT6 deacetylase activity after incubation of 1 μm WT hSIRT6 (●) or C18S mutant (○) with 20 μm OA, NO₂-OA, CLA, or NO₂-CLA. Error bars represent S.D. of three replicates. ****, p < 0.0001 ANOVA test.

Figure 7

In silico docking experiments.A, best poses after docking OA and NO₂-OA into the active site of hSIRT6. The carbon backbone is colored according to the docked molecule. Oxygen and nitrogen atoms are red and blue, respectively, whereas unsaturations are shown in gold. The co-substrate is represented with brown sticks. Relevant residues in the protein are pointed out. The molecular surface of the active site is shown in gray, whereas the reminding structural context of the protein is shown as a transparent cartoon. B, same as panel A for CLA and NO₂-CLA. C, best poses after docking OA and NO₂-OA considering the presence of an acetylated H3 peptide at Lys-9 (H3K9Ac). D, same as panel C for CLA and NO₂-CLA.

Figure 8

Molecular modeling of nitro-fatty acid conjugation to Cys-18.A, molecular representation of one possible conformation of the Michael adduct between Cys-18 and 9-NO₂-OA. The nitro-fatty acid is shown in balls and stick, whereas nearby protein residues are drawn in sticks. Residues that may form an exposed hydrophobic patch at the surface of the acyl/co-substrate–binding loop are labeled in magenta. The location and distance to the C-pocket is indicated. B, same as panel A but showing other alternative conformation that the same adduct may adopt.

Figure 9

In situ covalent modification of hSIRT6 by NO₂-OA promotes H3K9 deacetylation in cells.A, HEK293T cells expressing FLAG-SIRT6 were treated with 10 μm Bt-NO₂-OA for 1 h. After cell lysis, Bt-NO₂-OA adducts were affinity-purified using streptavidin-agarose beads followed by FLAG immunodetection (lanes 1–4). FLAG-SIRT6 from the same input lysates used for affinity purification is shown on the right (lanes 5-8). Immunoprecipitation was performed in HEK293T cells overexpressing FLAG-SIRT6 (lanes 3 and 4) and in control HEK293T cells without transfection (lanes 1 and 2). Arrows indicate 55- and 36-kDa molecular mass standards. B, HEK293T cells were treated with 10 μm OA or 10 μm NO₂-OA for 1 h in serum-free DMEM and H3K9 acetylation assessed by Western blotting analysis. C, densitometric H3K9Ac quantification versus total amount of H3 (H3K9Ac/H3). Statistically significant differences were observed between control and NO₂-OA incubation. Error bars represent S.D. of at least three replicates. **, p < 0.001 as determined by ANOVA.