Aryl Fluorosulfate Based Inhibitors That Covalently Target the SIRT5 Lysine Deacylase

Abstract

The sirtuin enzymes are a family of lysine deacylases that regulate gene transcription and metabolism. Sirtuin 5 (SIRT5) hydrolyzes malonyl, succinyl, and glutaryl ϵ-N-carboxyacyllysine posttranslational modifications and has recently emerged as a vulnerability in certain cancers. However, chemical probes to illuminate its potential as a pharmacological target have been lacking. Here we report the harnessing of aryl fluorosulfate-based electrophiles as an avenue to furnish covalent inhibitors that target SIRT5. Alkyne-tagged affinity-labeling agents recognize and capture overexpressed SIRT5 in cultured HEK293T cells and can label SIRT5 in the hearts of mice upon intravenous injection of the compound. This work demonstrates the utility of aryl fluorosulfate electrophiles for targeting of SIRT5 and suggests this as a means for the development of potential covalent drug candidates. It is our hope that these results will serve as inspiration for future studies investigating SIRT5 and general sirtuin biology in the mitochondria.

Keywords: Covalent Inhibition; Enzyme Inhibitors; SIRT5; Sirtuins; SuFEx.

Figure 1

First generation SIRT5‐targeting, aryl fluorosulfate‐based inhibitors. A) Structure of an analogue of 1 bound to SIRT5 (PDB 6EQS). B) Predicted binding mode of 4 (Schrödinger, Maestro suite), displaying close proximity of the electrophilic sulfur(VI) atom to the Tyr102 side chain. C) Chemical structures of the reversible SIRT5 inhibitor 1 and analogues containing various electrophilic warheads (2–8). D) Chemical structures containing an alkyne for click chemistry, combined with various electrophilic warheads (9–14).

Figure 2

Second generation SIRT5‐targeting, aryl fluorosulfate‐containing inhibitors. A) Structures of compounds 15–19. B) LC‐MS analysis of the time‐dependent formation of covalent conjugates between non‐tagged recombinant SIRT5 (10 μM) and compounds 16–19 (100 μM) in the presence of NAD⁺ (200 μM). *Corresponds to a byproduct where SIRT5 has been modified twice by adduct formation. C) Jump dilution assay performed for 17 and 19 after 16 h pre‐incubation. D), E) Determination of k _obs from time‐dependent dose‐response experiments and subsequent data fitting to k _obs=(k _inact×[I])/(K _I+[I]), to derive k _ianct and K _I values for 17 and 19, respectively. For kinetic model, additional equations and plots of the time‐dependent inhibition data, see Supporting Figure S10).

Figure 3

Selectivity for SIRT5 and targeting of its substrate binding pocket. A) Covalent labeling of SIRT1‐7 by 17. Recombinant enzymes (2.5 μM) were incubated with 17 and NAD⁺ (200 μM) (all enzymes except SIRT4 were applied in functional assays and shown to be active deacylases; see Supporting Figure S11 for full gel images, repetitions, and similar evaluation of 19). B) Labeling of non‐tagged recombinant SIRT5 by 17 and 19, with or without pre‐boiling or SDS treatment. For full gel images and repetitions see Supporting Figure S12. C) Concentration‐dependent competition of the covalent labeling of non‐tagged recombinant SIRT5 by 17, using the reversible inhibitor 1 as competitor. D) Concentration‐dependent competition of the covalent labeling of non‐tagged recombinant SIRT5 by 17, using the fluorogenic substrate Ac‐LGKglut‐AMC as competitor. For compound 19, full gel images, and repetitions related to (C) and (D), see Supporting Figure S13.

Figure 4

Labeling of SIRT5 wild‐type and active site mutants. A) LC‐MS analysis of the time‐dependent formation of covalent conjugates between recombinant SIRT5(Y102F) and SIRT5(R105A) mutants (10 μM) with compound 17 or 19 (100 μM) in the presence of NAD⁺ (additional data, including that for compounds 16, 18, and 19 and data for the mutants SIRT5(Y104F) and SIRT5(Y102F/Y104F) are available in the Supporting Information and Supporting Figure S9). *Corresponds to a biproduct where SIRT5 has been modified twice by adduct formation. B) Competition of covalent binding of compound 17 to the mutant enzymes by inhibitor 1, visualized by in‐gel fluorescence. For full gel images and repetitions, see Supporting Figure S15. C) LC‐MS/MS analysis of tryptic digests of SIRT5 or SIRT5(Y102F) after incubation with compound 16. Loss of peptides, VWEFY₁₀₂HYR, VWEFFHY₁₀₄R, and GAGGY₇₆WR, due to modification by the covalent inhibitor in the presence or absence of NAD⁺. Bars represent the fold change (%) normalized to unmodified samples. D) Representative MS/MS spectrum of the unmodified VWEFY₁₀₂HYR peptide and mirror of the modified VWEFY₁₀₂[+643.16 Da]HYR peptide after tryptic digestion of recombinant SIRT5. Modification was clearly identified as occurring at Tyr102. The presented m/z values are deconvoluted to a +1 charge state for ease of comparison (y5, y6, and y7 were observed as doubly‐charged fragments in the raw MS/MS spectrum, see Supporting Information). E) Representative MS/MS spectrum of the unmodified GAGGY₇₆WR peptide and mirror of the modified GAGGY₁₀₂[+643.16 Da]WR peptide after tryptic digestion of recombinant SIRT5. Modification was clearly identified at Tyr76. For data regarding modification of SIRT5(Y102F) on VWEFFHY₁₀₄RR, additional data, and repetitions, see Supporting Figures S16–S18.

Figure 5

Chemical stability of compounds 16 and 17. A) Stability in HDAC assay buffer at 37 °C measured by HPLC. B) Stability in HDAC assay buffer, containing reduced glutathione (GSH; 2 mM) at 37 °C. Data are shown as mean values relative to the arbitrary fluoresence units at t=0 h±SD (n=2). For additional data with compounds 18 and 19, see Supporting Figure S18). C) Structures of synthesized chloroalkane probes. D) CAPA results for compounds 20 and 21 as well as the α‐N‐“chloroalkane”‐containing tryptophan (CA‐Trp‐OH), compound 1 (CA‐1), and compound 1‐Et (CA‐1‐Et) after 4 hours of treatment with inhibitor (n≥3). See the Supporting Information Figure S19 for structures and synthesis of control CAPA compounds.

Figure 6

Targeting SIRT5 in cell lysates and living cells. A) Dose‐dependent pull‐down of overexpressed SIRT5‐FLAG in HEK293T cell lysate by compounds 17 and 19. Bar graphs represent % chemiluminescence signal, normalized to the band for 20 μM treatment (n=3; for full gel images, replicates and conditions, see Supporting Figure S22). B) Competition of the covalent conjugate formation with overexpressed SIRT5‐FLAG in HEK293T cell lysate (with 17 or 19 at 20 μM) by cotreatment with 1. Bar graphs represent % chemiluminescence signal, normalized to the band for 20 μM treatment with 17 or 19 (n=3; for full gel images, replicates, and conditions, see Supporting Figure S23). C) Labeling and competition of the labeling of overexpressed SIRT5‐FLAG in cultured HEK293T cells with 17 (20 μM, 5 h treatment) by cotreatment with prodrugs 1‐Et or NRD167. D) Labeling and competition of the labeling, of overexpressed SIRT5‐FLAG in cultured HEK293T cells with 19 (20 μM, 5 h treatment) by cotreatment with prodrugs 1‐Et or NRD167. Experiments shown in (C), (D) were repeated twice with similar results; for structures of compounds, full gel images, replicates, and conditions, see Supporting Figure S24. Significance of the levels of pulled‐down SIRT5 were calculated using one‐way ANOVA and Turkey's multiple comparison tests. Adjusted p values are in comparison to treatment with DMSO (A) or covalent inhibitors 17 or 19 without competitor (B): ns denotes p>0.05, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 7

Compound stability and labeling of SIRT5 in vivo. A) Stability of compounds 1, 17, and 19 in human serum at 37 °C. Aliquots were collected at different time points, quenched with urea, and partitioned between trichloroacetic acid‐acetone (1 : 9) by centrifugation (14000 g). The amount of compound left in the supernatant was analyzed by HPLC. B) Pull‐down of labeled SIRT5 in hearts from mice injected with a single dose of 17 (i.v., 12 mg kg⁻¹) and sacrificed after 6 or 24 h. The lysates from hearts were subjected to click chemistry with biotin‐N₃, followed by enrichment using streptavidin beads, SDS‐PAGE, and western blot analysis. C) Examples of western blots of SIRT5 pull‐downs from harvested organs. For additional data and conditions, see Supporting Figure S27.