Fig 1.
Design and validation of a SIRT2—myristoyl-peptide HTRF binding assay.
(A) Schematic of our HTRF binding assay where His tagged-SIRT2 interacts with a terbium cryptate-labeled antibody and a fluorescein-labeled myristoyl-peptide. Displacement of the peptide from SIRT2 by a competitor then reduces FRET efficiency after terbium cryptate excitation. (B) Fluorescence intensity in HTRF binding assays using different concentrations of SIRT2 and FAM-myristoyl-H4K16 peptide (excitation/emission wavelengths = 330 nm/520 nm, normalized by emission at 620 nm) (see Materials and Methods). (C) Displacement of FAM-myristoyl-H4K16 peptide from SIRT2 in the HTRF binding assay by an identical peptide lacking the FAM-PEG4 group. (D) Displacement of FAM-myristoyl-H4K16 peptide from SIRT2 in the HTRF binding assay by the SIRT2 deacetylase and defatty-acylase inhibitor ascorbyl palmitate. (E) Displacement of FAM-myristoyl-H4K16 peptide from SIRT2 in the HTRF binding assay by the SIRT2 deacetylase inhibitor SirReal2.
Fig 2.
High-throughput screen of 9600 compounds for SIRT2 binding in HTRF assays and the identification of 8008–3660 (1) as a SIRT2 deacylase inhibitor.
(A) 9600 compounds at a single concentration of 10 μM were screened for their ability to displace FAM-myristoyl-H4K16 peptide from SIRT2. The percent inhibition of peptide binding caused by all 9600 compounds is shown in the left panel. The middle panel shows the data for test compounds that inhibited SIRT2 binding to FAM-myristoyl-H4K16 peptide by >40%. These 63 compounds were used in full dose-response binding assays in HTRF format to confirm binding and to determine an IC50 for peptide displacement (S5 Fig). The 9 ligands in the table in the right of panel A inhibited FAM-myristoyl-H4K16 peptide binding to SIRT2 with IC50 values less than 10 μM. See S6 Fig for the chemical structures of all 9 compounds. (B) Effect of 10 μM compound on the deacetylase activity of SIRT2. The enzyme activity in the presence of compound was compared to a control SIRT2 assay (Ctrl) where the enzyme was treated with DMSO alone. (C) Effect of 50 μM compound on the demyristoylase activity of SIRT2 compared to a control SIRT2 assay (Ctrl) where the enzyme was treated with DMSO alone. Assays in panels B and C contained 0.7% DMSO from dissolving and diluting ligands. (D) Chemical structure of the identified SIRT2 ligand 8008–3660 (1).
Fig 3.
SIRT2 binding and inhibition of SIRT2 deacylase activities by 1.
(A) Displacement of Cy3-myristoyl-H4K16 peptide from SIRT2 by 1 as measured by a reduction in Cy3 fluorescence. The IC50 value of 16 μM was used to calculate a Kd of 5 μM for the interaction of 1 with SIRT2 (see Materials and Methods). (B) 1 inhibited SIRT2 deacetylase activity with an IC50 of 7 μM. (C) 1 inhibited SIRT2 de-decanoylase activity with an IC50 of 8 μM. (D) 1 inhibited SIRT2 de-dodecanoylase activity with an IC50 of 39 μM. (E) 1 inhibited SIRT2 demyristoylase activity with an IC50 of 37 μM. See Materials and Methods for assay conditions.
Fig 4.
Inhibition of SIRT2 by 1 in cells.
(A) Immunofluorescence images of acetylated alpha-tubulin in A549 cells showing enhanced fluorescence in cells treated with 25 μM of SirReal2 (positive control) or 100 μM of 1. The media for the control contained 2% DMSO, and the scale bar is 10 μm. (B) Comparison of GFP fluorescence intensities from acetylated alpha-tubulin staining from the images in panel A. GFP fluorescence intensities from individual cells were measured and normalized to the DAPI intensities of the same cells, and the intensities were compared with a one-way ANOVA followed by Dunnett’s multiple comparison test (***p < 0.001). (C) Western blot showing knockout of SIRT2 in A549 cells using CRISPR/Cas9. The alpha-tubulin blot, which was performed on a separate membrane, was used as a loading control. (D) Immunofluorescence images of A549 SIRT2-KO cells showing increased acetylated alpha-tubulin levels in cells without the protein, and the ligands SirReal2 and 1 had no effect on the levels of acetylated alpha-tubulin in SIRT2-KO cells. Experiments were performed identical to panel A. (E) Comparison of GFP fluorescence intensities from acetylated alpha-tubulin staining from the images in panel D. These were quantified and compared as described above for panel B. (F) SirReal2 treatment for 72 hours did not cause toxicity in A549 cells, as measured with an MTT assay. The control (Ctrl) cells were treated with 0.7% DMSO to be consistent with the drug treated cells. (G) Treatment of A549 cells with 1 for 72 hours did not cause toxicity, which was determined with an MTT assay as in panel F.
Fig 5.
The ability of 15 small molecules to bind SIRT2 or affect its deacetylase or demyristoylase activities.
“N.D.” means not determined, and “no effect” indicates that 200 μM compound did not significantly affect SIRT2 activity. The Kd value for 1 was determined from a Cy3-myristoyl-H4K16 peptide binding competition assay. The Kd values for 2, 4, 5, and 6 were determined from a 1-aminoanthracene binding competition assay. *All data for 3 was previously published [12]. See Materials and Methods for experimental details and the names of the chemicals.
Fig 6.
Enhanced affinity of 2 for the SIRT2—decanoyl-peptide complex compared to SIRT2 alone.
2 was used to displace the fluorescent ligand 1-aminoanthracene from SIRT2’s hydrophobic tunnel which resulted in a reduction of 1-aminoanthracene fluorescence. The concentration of 1-aminoanthracene was 100 nM, the SIRT2 concentration was 4 μM, and when added, the concentration of decanoyl-H4K16 peptide was 10 μM. The IC50 values were used to calculate Kd values of 68 μM for the interaction of 2 with SIRT2 and 3 μM for the interaction of 2 with the SIRT2—decanoyl-peptide complex [12]. The Kd of 68 μM for its SIRT2 interaction was also reported in Fig 5.