Profiling Sirtuin Activity Using Copper-free Click Chemistry

Abstract

The mammalian sirtuins are a group of posttranslational modification enzymes that remove acyl modifications from lysine residues in an NAD⁺-dependent manner. Although initially proposed as histone deacetylases (HDACs), they are now known to target other cellular enzymes and proteins as well. Sirtuin-catalyzed simple amide hydrolysis has profound biological consequences including suppression of gene expression, promotion of DNA damage repair, and regulation of glucose and lipid metabolism. Human sirtuins have been intensively pursued by both academia and industry as potential therapeutic targets for the treatment of diseases such as cancer and neurodegeneration. To gain a better understanding of their roles in various cellular events, innovative chemical probes are highly sought after. This current study focuses on the development of activity-based chemical probes (ABPs) for the profiling of sirtuin activity in biological samples. Cyclooctyne-containing and azido-containing probes were synthesized to enable the subsequent copper-free “click” conjugation to either a fluorophore or biotin. The two groups of structurally related ABPs demonstrated different labeling efficiency and selectivity: the cyclooctyne-containing probes failed to label recombinant sirtuins to any appreciable level, while the azido-containing ABPs showed good isoform selectivity. The azido-containing ABPs were further analyzed for their ability to label an individual sirtuin isoform in protein mixtures and cell lysates. These biocompatible ABPs allow the study of dynamic cellular protein activity change to become possible.

Graphical Abstract

INTRODUCTION

Sirtuins are a group of “eraser” enzymes that are responsible for removing acyl groups from lysine residues. They were classified as “class III HDACs” when first discovered.^{1, 2} Unlike the class I and II HDACs which are Zn⁺-dependent enzymes,³ sirtuins exert their catalytic activity in an NAD⁺-dependent manner,^{4, 5} which integrates posttranslational modifications (PTMs) with energy homeostasis so elegantly. Later studies significantly broadened the substrate range of sirtuins as they not only target histones, but also other cellular proteins and enzymes.^{6, 7} It has been suggested that sirtuins be renamed as “lysine deacetylases” (KDACs). Most recently, many novel activities were discovered for sirtuins. In addition to removing covalently attached acetyl groups, they can also “erase” propionyl, butyryl, malonyl, succinyl, and long-chain fatty-acyl groups from lysines.^8–12 The physiological significance of these novel PTMs is still under investigation.

The human sirtuin family comprises seven members, SIRT1-SIRT7.¹³ They demonstrate unique enzymatic activities, subcellular localizations, and biological functions.^14–16 Through deacylating various cellular targets, human sirtuins regulate a wide range of biological events including transcription silencing,⁷ mitochondrial biogenesis,¹⁷ and metabolism,¹⁸ among many others. SIRTs have been closely associated with various pathological conditions,^19–22 and have been implicated as either promoters or suppressors of distinct cancers.^23–26

With tremendous efforts in pursuing human sirtuins as viable therapeutic targets come significant controversies in the sirtuin biology field,²⁷ many of which are due to the lack of knowledge about the intricate regulatory network between SIRTs and their endogenous targets and interacting partners. To reconcile these controversies, novel chemical tools are absolutely needed. The classical method to study protein function involves the recombinant expression and purification of the protein target, followed by characterization using in vitro assays. Although useful, it is only applicable to proteins that can be expressed recombinantly. Furthermore, some proteins exist in large protein complexes endogenously. Upon isolation, they may lose activity entirely. And for some proteins, activity assays are not readily available. Proteomic analysis has been increasingly recognized as a powerful approach for functional annotations on a large scale.²⁸ In cells, the protein function can be regulated by several factors such as PTMs, endogenous protein binding partners, and endogenous inhibitors.²⁹ It is critical to point out that these above-mentioned factors may alter protein function/activity without changing protein abundance. Thus, abundance-based proteomic studies may only provide indirect information about protein function.

Recent studies suggested that activity-based chemical probes (ABPs) are capable of reporting the functional state of enzymes directly in a complex native matrix.^{30, 31} A typical ABP comprises a “warhead”, a tag, and a linker. “Warheads” are usually small molecules targeting the active site of an enzyme to form covalent adducts with the amino acid residues (AAs) in the active site. Most ABPs use electrophilic “warheads” to conjugate to nucleophilic AAs.²⁹ In the case when nucleophilic AAs are not available, photoaffinity groups such as benzophenone, diazirine or azide can be incorporated into the probe. Upon UV irradiation, these photoactivatable groups can form covalent bonds with the AAs in their vicinity. The tag can be a fluorophore for visualization or biotin for affinity enrichment.²⁹ Special attention is given to the size of the tags because large tags may interfere with the cell permeability and cellular distribution of the ABPs. Alternatively, “click” chemistry can be used to append a tag to the ABP after the photoaffinity labeling is complete.^{32, 33} The linker of an ABP tethers the “warhead” with the tag. The presence of the linker reduces the potential steric hindrance from the tags. The linker length and composition can be varied to sample the adequate proximity, favorable solubility and cell permeability.²⁹

Our group has developed two generations of ABPs for human sirtuins.^{34, 35} They are thioacyllysine peptide-based probes³⁶ with either a benzophenone or a diazirine photoaffinity group, and a terminal alkyne “clickable” tag. They demonstrated good labeling efficiency and isoform selectivity in labeling individual recombinant sirtuins, and one or more isotypes in protein mixtures or cell lysates. However, the cross conjugation of these ABPs with a fluorophore or biotin requires Cu (I) catalyst. The cytotoxicity of copper prevents their applications in live cell studies.^{37, 38} The Cu-free renditions of these probes are needed.

Herein, we report the development of cyclooctyne- and azido-containing ABPs for human sirtuins (Fig. 1). These ABPs inherited the peptide backbone, the thioacyl “warheads”, and the benzophenone photoaffinity group from the two prior generations. Instead of a terminal triple bond, they featured a cyclooctyne or an azido tag to enable the subsequent Cu-free cycloaddition to a fluorescent dye or biotin. The cyclooctyne-based probes were proven to be sirtuin inhibitors with micromolar potencies. However, they showed negligible levels of labeling of the recombinant proteins. The azido-containing probes, in contrast, were selective SIRT2 inhibitors with IC₅₀ values in the low micromolar range. More importantly, they demonstrated robust and isoform selective labeling of SIRT2 in purified proteins and complex proteomes. The on-target effect was confirmed by a competition analysis with a known SIRT2 inhibitor. And these probes only labeled wildtype SIRT2 (wtSIRT2), but not its catalytically inactive mutant. Furthermore, the cellular sirtuin activity change in response to pharmacological interventions can also be captured and reported by these ABPs. Taken together, our study provide more powerful biocompatible chemical tools to investigate sirtuin activity change in native biological milieu.

Figure 1.

Chemical structures of ABPs.

METHODS AND MATERIALS

Reagents and Instruments

All reagents were purchased from Aldrich or Fisher Scientific and were of the highest purity commercially available. UV spectra were obtained with a Varian Cary 300 Bio UV-visible spectrophotometer. HPLC was performed on a Dionex Ultimate 3000 HPLC system equipped with a diode array detector using Macherey-Nagel C18 reverse-phase column. NMR spectra were acquired on a Bruker AVANCE III 500 MHz high-field NMR spectrometer and the data were processed using Topspin software. HRMS spectra were acquired with either a Waters Micromass Q-tof Ultima or a Thermo Scientific Q-Exactive hybrid Quadrupole Orbitrap. Fluorescence scanning was performed on a Biorad Versa Doc 4000 MP Imaging System or a Biorad ChemiDoc MP imaging system.

Synthetic Peptides

Synthetic peptides H3K9Ac: ARTKQTAR(K-Ac)STGGKAPRKQLAS, p53K382Ac: KKGQSTSRHK(K-Ac)LMFKTEG were synthesized and purified by Genscript. The peptides were purified by HPLC to a purity >95%.

Protein Expression and Purification

Plasmids of SIRT1 (full length), SIRT2 (38–356), SIRT3 (102–399), SIRT5 (34–302) and SIRT6 (1–314) were generous gifts from Dr. Hening Lin (Cornell University). The proteins were expressed and purified according to previously published protocols.³⁹ The identity of the protein was confirmed by tryptic digestion followed by LC-MS/MS analysis performed at the Vermont Genetic Network (VGN) Proteomics Facility. Protein concentrations were determined by Bradford assay.

Sirtuin Inhibition Assay

A typical reaction contained 500 μM NAD⁺, 500 μM peptide substrate (H3K9Ac for SIRT2, SIRT3 and SIRT6, p53K382Ac for SIRT1 and SIRT5), varying concentrations of small molecule probe in 100 mM phosphate buffer pH 7.5. The reactions were initiated by the addition of 10 μM of sirtuin and were incubated at 37°C before being quenched by 8 μL of 10% TFA. The incubation time was controlled to achieve less than 15% substrate conversion (10 min for SIRT1, 5 min for SIRT2, 15 min for SIRT3, 45 min for SIRT5, and 2 h for SIRT6). The samples were then injected on an HPLC fitted to a Macherey-Nagel Nucleosil C18 column. NAD⁺, NAM and AADPR peaks were resolved using a gradient of 0 to 20% methanol in 20 mM ammonium acetate. Chromatograms were analyzed at 260 nm. Reactions were quantified by integrating areas of peaks corresponding to NAD⁺ and AADPR. Rates were plotted as a function of small molecule probe concentration, and points were fitted to the following equation:

$$\nu (\%)={\nu }_0(\%)-\left[{\nu }_0(\%)\left({10}^{\text{x}}\right)/\left({10}^{\text{x}}+{\text{IC}}_{\text{50}}\right)\right]$$

where ν(%) represents turnover rate expressed as percent enzymatic activity remaining, ν₀(%) represents the uninhibited turnover rate expressed as an enzymatic activity of 100%. The variable x represents the log[probe] in nanomolar. IC₅₀ values were derived from this equation.

Labeling of Recombinant Sirtuin

For cyclooctyne-containing probes 1 and 2, a typical labelling experiment was performed as follows: in a 0.7 mL Eppendorf tube, purified recombinant human sirtuin (10 μM) was incubated with NAD⁺ (500 μM) and activity-based probe at 37 °C for 10 min. The sample was transferred to a clear-bottom 96-well plate, placed on ice, and irradiated at 365 nm with a UV-lamp in a cold room. Subsequently, TAMRA-azide (Click Chemistry Tools) was added, and the sample was gently agitated at 250 rpm on a microshaker at room temperature for 30 min. Then, the sample was resolved by SDS-PAGE. To reduce the signal to noise ratio, the gel was destained to eliminate non-specific binding of free dyes. This was done in a mixture of methanol/distilled water/acetic acid (v/v/v = 4/5/1) at ambient temperature for 4 h. The destained gel was analyzed with in-gel fluorescence scanning using a Biorad ChemiDoc MP imager (excitation 532 nm, 580 nm cut-off filter and 30 nm band-pass). Finally, Coomassie blue staining was applied to provide loading control.

For azido-containing probes 3 and 4, after the initial incubation and photoaffinity labeling, the samples were incubated with 15 mM iodoacetamide (IA) for 30 min at room temperature before TAMRA-DBCO was introduced. The subsequent steps were the same as described above.

Cell Culture

HEK293 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin and 100 mg/mL streptomycin. Cells were maintained in a humidified 37°C incubator with 5% CO₂.

Overexpression of SIRT2 in HEK293 Cells.

SIRT2 Flag was a gift from Eric Verdin (Addgene plasmid #13813).⁴⁰ The vector was transfected into HEK293 cells with lipofectamine 2000 (Thermo Fisher Scientific) following the manufacturer’s protocol. Overexpression efficiency was determined with western blot.

Cell Lysate Labeling

Cells were harvested and lysed with lysis buffer (50 mM Tris-HCl, pH 7.5, 300 mM NaCl, 1% Triton X-100) supplemented with protease inhibitor cocktail (Thermo Fisher Scientific). Protein concentration was determined by Bradford assay. A typical labeling experiment contained 10 to 50 μg protein, 500 μM NAD⁺, and 10 μM probe in 50 mM Tris-HCl buffer pH 7.5. The photoaffinity labeling, SPAAC-mediated conjugation to a fluorescent dye and visualization were similar to the protocols for labeling recombinant sirtuins as described above.

Sirtuin Enrichment in Cell lysate

HEK293 cell lysate was incubated with the chemical probe. After the UV irradiation, the lysate was treated with 15 mM IA for 30 min at room temperature. Subsequently, the sample was conjugated to cleavable diazo biotin-DBCO (Click Chemistry Tools). The biotinylated proteins were captured by high capacity streptavidin beads (Thermo Fisher Scientific). The captured protein can be eluted by incubation with 25 mM of Na₂S₂O₄, 250 mM of NH₄HCO₃ and 0.05% SDS for 1 h. The eluent was concentrated by lyophilization and analyzed by western blot.

Western Blot

The samples were resolved on a 10% SDS-PAGE gel and transferred to the Immun-Blot PVDF membrane (Biorad). The blot was blocked with 5% nonfat milk in TBST, probed with the primary antibody targeting SIRT2 (Cell Signaling Technology), washed with TBST, followed by incubation with the anti-rabbit HRP conjugated secondary antibody. The signal was then detected by Clarity^™ western ECL substrate (Biorad).

Cellular Imaging

Cells were seeded in a 4-well chamber slide (Thermo Fisher) and grew till 70 to 80% confluence. The cells were then incubated with either 50 μM probe 3 alone or a combination of 50 μM probe 3 and 20 μM AGK2 for 1 h before the chamber slide was placed on ice. The cells were irradiated at 365 nm in cold room for 30 min followed by fixation with 4% paraformaldehyde and permeabilization with 0.2% Triton X-100 in PBS. The cells were then treated with 15 mM IA for 30 min. Subsequently, the cells were incubated with 10 μM TAMRA-DBCO for 30 min at room temperature in the dark with gentle agitation. The cells were rinsed with PBS three times. Afterwards, the cells were incubated with anti-SIRT2 antibody (Cell Signaling Technology) at room temperature for 1 h, washed three times with 0.05% Tween-20 in PBS, and then incubated with anti-rabbit Alexa Fluor^™ Plus 488 IgG antibody (Invitrogen) for 30 min at room temperature. Cells were rinsed again with 0.05% Tween-20 in PBS three times with gentle agitation. For nuclear DNA staining, the cells were incubated with 3 μM DAPI in PBS for 10 min at room temperature. Cells were rinsed with PBS three times and deionized water once before images were acquired.

Fluorescence microscopy was carried out on a Leica TCS SP5 confocal microscope (Leica Microsystems, Germany). Alexa Fluor 488 was excited at 488 nm, and emission was obtained at 495–525 nm. TAMRA-DBCO was excited at 543 nm, and fluorescent emission was obtained at 565–595 nm. Images were acquired and processed using Leica Application Suite Advanced Fluorescence (LAS AF) software.

RESULTS AND DISCUSSION

Synthesis and characterization of cyclooctyne- and azido-containing ABPs

The synthetic scheme of cyclooctyne-containing probes was illustrated in Figure 2. Intermediate 3 has been prepared in our lab before.³⁴ Silver perchlorate triggered ring opening of commercially available 9,9-dibromobicyclo(6.1.0)nonane resulted in the production of a carbocation intermediate, which can be captured by ethylene glycol.⁴¹ Alkaline elimination and subsequent Mitsunobu reaction with intermediate 3 and saponification led to the formation of intermediate 5. Pre-assembled tripeptides 9 and 12 bearing either a thioacetyl or a thiomyristoyl warhead can then be conjugated to intermediate 5 using standard peptide bond formation condition to generate probes 1 and 2. All the new compounds were fully characterized by ¹H NMR, ¹³C NMR and HRMS.

Figure 2.

Synthesis of cyclooctyne-containing ABPs. a. AgClO₄, acetone (40%); b. DBU, DMSO, 60°C (55%); c. Intermediate 3, DIAD, TPP, THF, 0°C (54%); d. LiOH, THF/H₂O (81%); e. 5% Na₂CO₃/MeOH (54% for intermediate 8, 75% for intermediate 11); f. 20% piperidine, DMF (quantitatively); g. Intermediate 5, PyBOP, DIPEA, THF (44% for probe 1, 82% for probe 2).

The synthesis of azido-containing probes started with the mono-substitution of (E)-1,4-dibromobut-2-ene as described before (Figure 3).⁴² This intermediate was further elaborated by two more steps to form intermediate 15: etherification with intermediate 3, and deprotection of methyl ester by lithium hydroxide in a mixture of dioxane and water. Routine amide formation between intermediates 15 and 9 afforded probe 3 in 75% yield. Similarly, the coupling between intermediates 15 and 12 resulted in the formation of probe 4 in 77% yield.

Figure 3.

Synthesis of azido-containing ABPs. a. NaN₃, DMF, 50°C (50%); b. Intermediate 3, K₂CO₃, DMF, 60°C (61%); c. LiOH, THF/H₂O (quantitatively); d. Intermediate 9 or 12, PyBOP, DIPEA, THF (75% for probe 3, 77% for probe 4).

Biochemical characterization of ABPs

Inhibition of recombinant human sirtuins by the probes were assessed using an HPLC-based assay as described before.^{34, 35} The thioacetyl probe 1 was a SIRT1-selective inhibitor with an IC₅₀ of 3.00 ± 0.09 μM (Table 1, Figure S1). This probe also inhibited SIRT2 and SIRT3 with IC₅₀ values in the mid-micromolar range (82.2 ± 7.68 μM for SIRT2, and 65.3 ± 10.5 μM for SIRT3, respectively). It demonstrated 27- and 22-fold preference against SIRT1 over SIRT2 and SIRT3, respectively. The thiomyristoyl probe 2, on the other hand, only showed mild inhibition against SIRT2 with the IC₅₀ value of 415.8 ± 65 μM. Previous studies from our lab have suggested that thioacetyl ABPs were pan-human sirtuin inhibitors.^{34, 35} They demonstrated broader inhibitions against human sirtuin isoforms compared to their thiomyristoyl counterparts.^{34, 35} The selectivities of probe 1 and 2 were consistent with these prior observations.

Table 1.

IC₅₀ values of the ABPs.

Sirtuin	Substrate	IC50 (μM)
		1	2	3	4
SIRT1	p53K382Ac	3.00 ± 0.09	NI	NI	NI
SIRT2	H3K9Ac	82.2 ± 7.68	415.8 ± 65	2.98 ± 0.04	9.14 ± 0.98
SIRT3	H3K9Ac	65.3 ± 10.5	NI	35.9 ± 5.82	303.1 ± 52.5
SIRT5	p53K382Ac	NI	NI	NI	NI
SIRT6	H3K9Ac	NI	NI	NI	618.7 ± 49

Probe 3, with the thioacetyl warhead and azido bioorthogonal tag, selectively inhibited SIRT2 with an IC₅₀ of 2.98 ± 0.04 μM, a 12-fold selectivity over SIRT3 (Table 1, Figure S1). It failed to inhibit SIRT1, SIRT5 and SIRT6 to any appreciable extent. Comparing to the terminal alkyne-containing ABPs with the same thioacetyl warhead reported previously,^{34, 35} probe 3 demonstrated improved isoform-selectivity and potency. Probe 4, the thiomyristoyl analog of probe 3, also showed isoform selectivity. It inhibited SIRT2, SIRT3 and SIRT6 with IC₅₀ values of 9.14 ± 0.98 μM, 303.1 ± 52.5 μM, and 618.7 ± 49 μM, respectively (Table 1, Figure S1), consistent with the notion that these three sirtuin isoforms also harbor defatty-acylase activity.¹⁰ Even at high micromolar concentrations, probe 4 did not inhibit SIRT1 or SIRT5. The SIRT2-selectivity of probe 4 outperformed the other thiomyristoyl ABPs we generated before,^{34, 35}

Recombinant human sirtuin labeling using cyclooctyne- or azido-containing ABPs

Photoaffinity labeling of recombinant sirtuins with the synthetic probes serves as the model study. Recombinant protein has the advantage of being ample in amount and free of background noise. Important labeling parameters can be easily optimized to enable the accurate profiling of active enzymes. For cyclooctyne-containing probes 1 and 2, the enzyme was incubated with NAD⁺ and the probe, followed by UV-irradiation at 365 nm for 1 h. The samples were then conjugated to TAMRA-azide in a copper-free fashion as detailed in “Methods and Materials”. Finally, the samples were resolved by SDS-PAGE and analyzed by in-gel fluorescence scanning. Surprisingly, although probes 1 and 2 both demonstrated inhibitory effects against human sirtuins in the above-mentioned in vitro assays, they both failed to label the recombinant sirtuins to any detectable levels. We reasoned that the bulkiness of the cyclooctyne group may have created steric hindrance to prevent the interaction between the photoaffinity group and the protein target. The fact that inhibition was detected but not the labeling suggested that target engagement using the thioacyl warhead still occurred. However, the distance or orientation of the photoaffinity group with the protein target was distorted by the large cyclooctyne moiety, leading to unsuccessful labeling.

Azido-containing probes 3 and 4 were also assessed for their ability to label human sirtuin isoforms using a similar approach as described above. TAMRA-DBCO (Figure 4A) was chosen as the fluorophore in the copper-free “click” conjugation for its fast kinetics and good stability under physiologically relevant conditions. Since both probes 3 and 4 showed preference towards human SIRT2 in the activity assessment, recombinant SIRT2 was used for the initial labeling experiments. Following the incubation and photocrosslinking as described above, TAMRA-DBCO was introduced into the sample for the strain-promoted alkyne-azide cycloaddition^{43, 44} (SPAAC, Figure 4B). Probe 3 was able to label SIRT2 in a concentration-dependent manner. However, background signal was detected in the negative control (Figure 5A, left), likely due to the spontaneous thiol-yne conjugation between the sulfhydryl groups of the cysteine residues and the highly constrained DBCO moiety.⁴⁵ Indeed, this undesired side reaction can be prevented by incubating the protein with excess amount of iodoacetamide (IA) to mask the reactive thiols^{45, 46} before the addition of TAMRA-DBCO (Figure 5A, right).

Figure 4.

Cu-free “click” labeling of sirtuins. A. Chemical structure of TAMRA-DBCO; B. Schematic representation of the photoaffinity and Cu-free “click” labeling of human sirtuins.

Figure 5.

SPAAC-mediated labeling of human sirtuins using azido-containing probe 3. A. Concentration-dependent labeling of SIRT2 (10 μM) without (left) or with (right) iodoacetamide (IA) treatment. The detailed labeling protocol was described in “Materials and Methods”; B. Labeling of recombinant SIRT3; C. Dose-dependent labeling of SIRT2 in mixtures of partially purified recombinant SIRT1, SIRT2, SIRT3, SIRT5 and SIRT6.

Probe 3 also demonstrated dose-dependent labeling of human SIRT3, albeit at higher concentrations (Figure 5B). The labeling patterns of SIRT2 and SIRT3 by this probe were consistent with the inhibitory effects on these two sirtuin isoforms. Furthermore, when probe 3 was incubated with a mixture of several partially purified human sirtuins (SIRT1, SIRT2, SIRT3, SIRT5 and SIRT6), SIRT2 was the only isoform that can be robustly labeled in a concentration-dependent fashion (Figure 5C). This remarkably simple chemical probe exhibited strong isoform selectivity in labeling recombinant sirtuins.

Probe 4 features a thiomyristoyl “warhead” which can be recognized by multiple sirtuin isoforms (Table 1). It showed strong preference to SIRT2, a 33-fold and a 67-fold selectivity over SIRT3 and SIRT6, respectively. As expected, probe 4 showed robust labeling of recombinant wtSIRT2 (Figure 6A, left). In contrast, it failed to label SIRT2H187Y, a catalytically inactive mutant, to any appreciable levels (Figure 6A, right). The sensitivity of probe 4 was superb as it labeled SIRT2 even in the nanomolar range (Figure 6B). SIRT3 and SIRT6 were the only other sirtuin isotypes labeled by probe 4, although at high micromolar concentrations (Figure S2). The labeling of SIRT2 in sirtuin mixtures further demonstrated the selectivity of this probe (Figure 6D). This labeling was out-competed (Figure 6D, far right lane) with the addition of excess amount of probe 5 (Figure 6C), a known non-clickable SIRT2 inhibitor,^{34, 47} suggesting the on-target effect of probe 4.

Figure 6.

Activity-based protein labeling using probe 4. A. Probe 4 robustly labeled wtSIRT2 (left), but failed to label the catalytically inactive mutant, SIRT2H187Y (right); B. Probe 4 labeled SIRT2 at nanomolar concentrations; C. Chemical structure of non-clickable thioacetyllysine tripeptide, probe 5; D. Selective labeling of SIRT2 in sirtuin mixtures. The labeling was blocked with the addition of excess amount of probe 5.

The cyclooctyne-containing probes showed inhibitory activities against several sirtuin isoforms. However, the steric bulkiness of the cyclooctyne moiety prevented them from labeling target proteins effectively. The azido-bearing probes were potent SIRT2 inhibitors. They exhibited good labeling efficiency, isoform selectivity, and sensitivity.

Cell lysate labeling and live cell imaging using azido-containing ABPs

The robust and concentration-dependent labeling of recombinant SIRT2 by azido-containing ABPs encouraged us to evaluate the labeling of this sirtuin isoform in a complex biological sample such as cell lysate. Rosetta^™(DE3) competent cells were transfected with the pET28a-SIRT2 plasmid. The cells were cultured and lysed as described in “Methods and Materials”. The E. coli cell lysate was used for the labeling experiments without further purification. The initial labeling focused on varying the lysate concentrations. As shown in Figure 7A, probe 3 selectively labeled SIRT2 with only 12 μg cell lysate. Optimal labeling was detected using 24 μg cell lysate. Therefore, 24 μg cell lysate was used for the subsequent labeling experiments. The cell lysate was incubated with NAD⁺ and varying concentrations of probe 3 before being subjected to the photoaffinity labeling and SPAAC-mediated conjugation to the fluorophore. A clean concentration-dependent labeling of SIRT2 was observed with increasing dosage of probe 3 (Figure 7B).

Figure 7.

Cell lysate labeling using azido-containing ABPs. A. Probe 3 selectively labeled SIRT2 in SIRT2-overexpressing E. coli cell lysate; B. Concentration-dependent labeling of SIRT2 by probe 3 in SIRT2-overexpressing E. coli cell lysate; C. SIRT2 was labeled in SIRT2-overexpressing HEK293 cell lysate. With 50 μg cell lysate, other cellular proteins were also labeled; D. SIRT2 labeling and enrichment. SIRT2-overexpressing HEK293 cell lysate was incubated with probe 3, followed by UV-irradiation. The sample was treated with excess amount of IA before conjugation to diazo biotin-DBCO. The biotinylated proteins were captured by streptavidin beads. The enriched proteins can then be cleaved off using Na₂S₂O₄. The identities of the eluted proteins were confirmed by western blot using an anti-SIRT2 antibody; E. Endogenous protein labeling using probe 3.

Probe 3 not only labeled SIRT2 in E. coli cell lysates, but also mammalian cell lysates. HEK293 cells were transfected to overexpress SIRT2. The lysates were treated with 10 μM probe 3. The labeling of SIRT2 started to be detected with 20 μg cell lysates (Figure 7C), and intensified with increasing amounts of lysates. With 50 μg cell lysates, other cellular proteins were also labeled (Figure 7C, far right lane). The azido moiety in probe 3 not only allowed the labeled protein to be conjugated to a fluorophore for visualization, but also to biotin for enrichment. In this case, after the photocrosslinking and IA treatment, the sample was incubated with the cleavable diazo biotin-DBCO. The biotinylated proteins can then be enriched by streptavidin beads and cleaved off the beads with Na₂S₂O₄. The eluted proteins were then resolved by SDS-PAGE and analyzed with an anti-SIRT2 antibody (Figure 7D). The enrichment experiment further confirmed that the protein being labeled in these complex proteomes was indeed SIRT2.

Endogenous protein labeling was also performed using probe 3. HEK293 cell lysates were treated with 0, 10 or 25 μM probe 3. Endogenous SIRT2 was labeled in a concentration-dependent manner (Figure 7E). With increasing concentrations of probe 3, the labeling of other proteins was detected as well. Probe 4 demonstrated comparable labeling capacity in cell lysate samples (Figure S3).

The superb performance of probe 3 in the aforementioned studies prompted us to conduct live cell labeling and imaging using this ABP. HEK293 cells were transfected to transiently overexpress human SIRT2. The cells were incubated with 50 μM probe 3. For comparison, another set of cells were cultured with a combination of 50 μM probe 3 and 20 μM AGK2,⁴⁸ a known SIRT2 inhibitor. The cells were then irradiated at 365 nm for 30 min on ice. After IA treatment, Cu-free “click” conjugation to TAMRA-DBCO was performed, followed by the incubation with anti-SIRT2 antibody and DAPI before the images were acquired. Pharmacological inhibition of SIRT2 by AGK2 significantly reduced the labeling intensity (Figure 8, TAMRA-DBCO) without changing SIRT2 protein levels (Figure 8, SIRT2). These results suggested that probe 3 can measure relative SIRT2 activity change in response to pharmacological modulations independent of protein abundance.

Figure 8.

In-cell labeling of SIRT2 with probe 3. SIRT2-overexpressing HEK293 cells were treated with either probe 3 alone (50 μM) or a combination of probe 3 (50 μM) and AGK2 (20 μM). Subsequently, the cells were irradiated at 365 nm, followed by IA treatment before conjugation to TAMRA-DBCO. The cells were then incubated with anti-SIRT2 antibody and DAPI before images were acquired. Treatment with AGK2 attenuated the labeling of SIRT2 by probe 3 without changing SIRT2 protein level.

Probe 3 not only labeled recombinant SIRT2, but also showed selective labeling of this particular sirtuin isoform in protein mixtures as well as native proteomes. The labeling efficiency of this probe decreased in cell lysates as compared to recombinant proteins, likely due to the interaction of probe 3 with other cellular targets in the proteomic preparations. In the cellular studies, the dynamic change of SIRT2 activity upon AGK2 treatment was captured and accurately reported by probe 3.

CONCLUSIONS AND PERSPECTIVES

In the current study, our effort has been dedicated to the synthesis and characterization of cyclooctyne- and azido-containing ABPs to profile sirtuin activity using Cu-free “click” chemistry. These probes featured a simple Ala-Ala-Lys tripeptide backbone with a benzophenone photoactivatable moiety, and either a cyclooctyne or azido “clickable” tag. This group of ABPs were sirtuin inhibitors with distinct isoform selectivity and potency. Unfortunately, the cyclooctyne-containing probes 1 and 2 failed to label any recombinant sirtuins, mainly due to the steric hindrance of the eight-membered ring moiety. Probes 3 and 4, on the contrary, exhibited robust labeling of individual sirtuin isoforms, but not their catalytically inactive mutants. They selectively highlighted SIRT2 in the recombinant sirtuin mixtures. This labeling can be effectively blocked in the presence of a known SIRT2 inhibitor, indicating the on-target activity of these ABPs. The power of these ABPs was further demonstrated in the cell lysate labeling experiments. SIRT2 was selectively labeled by probe 3 in a concentration-dependent manner in both the SIRT2-overexpressing E. coli and HEK293 cell lysates. Additionally, probe 3 was capable of labeling endogenous SIRT2 in mammalian cell lysate, albeit with reduced sensitivity and selectivity.

These probes create new opportunities to directly measure the functional state of sirtuins in their native environment independent of their protein abundance. They are truly biocompatible ABPs, taking advantage of the Cu-free “click” reaction. They can be applied to cellular imaging studies to capture the dynamic change of protein activity and cellular localizations, as evidenced by our in-cell labeling studies. Further optimizations of these probes are needed to improve labeling efficiency, isoform selectivity, and target range. For example, alteration of the position and accessibility of the photoactivatable groups and “click” tags may effectively translate binding into covalent conjugation. Special attentions will also be given to the modification of the peptide backbone to achieve broader target range because some of the sirtuin isoforms have distinct substrate specificity.^{49, 50} The probes reported in the current study and their structural variants will greatly expand our toolbox for the functional investigation of sirtuin family of protein deacylases in the complex biological settings.