Trichostatin A Inhibits Deacetylation of Histone H3 and p53 by SIRT6

Abstract

SIRT6 is an epigenetic modification enzyme that regulates gene transcription through its deacetylase activity. In addition to histone protein, SIRT6 also modify other proteins and enzymes, some of which are central players in metabolic reprogramming and aging process. Therefore, SIRT6 has emerged as a therapeutic target for the treatment of metabolic disorder and age-related diseases. Here, we report that SIRT6 deacetylates lysine 382 of p53 in short synthetic peptide sequence and in full length p53. Further studies showed that the deacetylation of H3K9Ac and p53K382Ac are insensitive to nicotinamide inhibition, but are sensitive to trichostatin A (TSA) inhibition. Detailed kinetic analysis revealed that TSA competes with the peptide substrate for inhibition, and this inhibition is unique to SIRT6 in the sirtuin family. Taken together, this study not only suggests potential roles of SIRT6 in regulating apoptosis and stress resistance via direct deacetylation of p53, but also provides lead compound for the development of potent and selective SIRT6 inhibitors.

INTRODUCTION

The mammalian sirtuins are NAD⁺-dependent protein deacetylases that regulate chromatin, transcription factors, co-transcription regulators, cytoskeletal proteins and metabolic enzymes via their catalytic activities. In human there are seven sirtuin isoforms (SIRT1-7)^1,2 located within distinct subcellular compartments, with SIRT1, SIRT6 and SIRT7 found predominantly in the nucleus, SIRT2 in the cytosol, and SIRT3, SIRT4 and SIRT5 in the mitochondrion.² They play critical roles in apoptosis,^3,4 metabolism,^5,6 mitochondrial biogenesis,^7,8 DNA repair,⁹ insulin secretion,¹⁰ and neuroprotection.^11,12 Recent studies also imply that sirtuins are important cell adaptor proteins that respond to low calorie conditions with the alteration of cell physiology.

Among the “magnificent seven” human sirtuins, SIRT6 is of special interest. SIRT6 locates in the nuclear compartment, and is tightly associated with chromatin.¹³ The N-terminus of SIRT6 is indispensable for its enzymatic activity and association with chromatin, while the C-terminus is imperative for its subcellular localization.¹³ Recent studies have uncovered the critical role of SIRT6 in maintaining genomic integrity and regulating metabolic network.^{5,14,15,16,17} SIRT6 knockout mice demonstrated premature aging phenotype due to dysregulated metabolism and genome instability.^5,18 Overexpression of SIRT6, however, extends the lifespan of male mice with improved metabolic profiles as compared to the control littermates.¹⁷ Additionally SIRT6 has also been implied as a tumor suppressor.^19,20 At the molecular level, SIRT6 controls various cellular events through its deacetylation activity. For example, acetylated histone H3K9 was the first identified physiological substrate of SIRT6.¹⁴ In vitro screening of acetylated histone tail peptides revealed the specific deacetylation activity of SIRT6 against acetylated H3K9.¹⁴ In SIRT6-deficent cells hyperacetylation of H3K9 was detected at telomeres.¹⁴ This modification regulates the interaction between telomeric chromatin and target genes.¹⁴ SIRT6 depletion therefore compromises telomere integrity and drives cells entering premature senescence.¹⁴ In addition to the deacetylase activity, SIRT6 also possesses mono-ADP ribosylase²¹ and defatty-acylase activities.^22,23 Through these novel activities, SIRT6 promotes DNA repair,²¹ facilitates the secretion of tumor necrosis factor-α (TNF-α),²² and regulates the membrane localization of R-Ras2.²⁴

In cells the biochemical activity of SIRT6 is subjected to multiple forms of regulation. During calorie restriction (CR) or fasting, SIRT6 demonstrates increased gene expression and protein abundance.²⁵ This induction triggers the metabolic reprogramming and physiologic adaptation to nutritional cues.²⁵ On the other hand, intracellular NAD⁺ level serves as another critical control point of SIRT6 activity. Pharmacological inhibition of nicotinamide phosphoribosyltransferase (NAMPRT), the rate limiting enzyme in NAD⁺ biosynthetic pathway, led to the depletion of cellular NAD⁺ content.²⁶ Subsequently the production of TNF-α decreased, presumably due to SIRT6 activity reduction.²⁶ Accumulating data suggests that targeted inhibition of SIRT6 could serve as a novel therapeutic approach for a broad spectrum of diseases. But to our knowledge so far only a few SIRT6 inhibitors have been identified.^27,28,29 They have demonstrated modest potency and limited isoform selectivity.^27,28,29 The need for novel scaffolds that can selectively inhibit SIRT6 becomes apparent.

Herein, we report the discovery of a pharmacological SIRT6 inhibitor, trichostatin A (TSA). TSA, a known inhibitor of the canonical Class I and Class II histone deacetylases (HDACs), reduces SIRT6 catalyzed deacetylation of synthetic peptide substrates as well as full length histone proteins isolated from HEK293 cells. This inhibition is very specific for SIRT6 as other mammalian sirtuins were not inhibited by TSA. Additionally, we have identified that SIRT6 deacetylates p53 at lysine 382 in vitro. This deacetylation can also be inhibited by TSA. Tumor suppressor p53 plays critical roles in various signaling pathways, many of which regulate cell cycle and stress-induced apoptosis. Previously, p53 has been identified as an endogenous substrate of SIRT1. SIRT1 physically interacts with p53 and deacetylates lysine 382 in a NAD⁺-dependent fashion.³⁰ This modification attenuates p53 transcriptional activity and downregulates p53-mediated apoptosis during severe stress.^30,31 Recent study has demonstrated that p53 upregulates SIRT6 expression. Subsequently, the deacetylation of forkhead box protein O1 (FoxO1) by SIRT6 causes the nuclear exclusion of this transcription factor, leading to the down-regulation of several gluconeogenic genes.³² It has long been postulated that SIRT6, another nuclear sirtuin, may also regulate p53 activity through direct deacetylation.³³ Our results shed important light on the cellular target, biological function and regulation of this intriguing enzyme.

METHODS AND MATERIALS

Reagents and Instrument

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. Radiolabeled samples were counted in a Beckman LS6500 scintillation counter. HRMS was acquired with either a Q-Exactive Hybrid Quadrupole Orbitrap Mass Spectrometer (Thermo Scientific) coupled to an EASY-nLC 1000 Liquid Chromatography (Thermo Scientific), or a LTQ Orbitrap Discovery (Thermo Scientific) Mass Spectrometer coupled to a Surveyor HPLC system (Thermo Scientific).

Synthetic Peptides

Synthetic peptides H3K9Ac: ARTKQTAR(K-Ac)STGGKAPRKQLAS, p53K382Ac: KKGQSTSRHK(K-Ac)LMFKTEG, p53K381Ac: KKGQSTSRH(K-Ac)KLMFKTEG, p53K373Ac: K(K-Ac)GQSTSRHKKLMFKTEG, p53K372Ac: (K-Ac)KGQSTSRHKKLMFKTEG, p53K120Ac: LHSGTA(K-Ac)SVT 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 (full length) were the generous gifts from Dr. Hening Lin (Cornell University). The proteins were expressed and purified according to previously published protocols.³⁴ The identity of the protein was confirmed by tryptic digestion followed by LC-MS/MS analysis performed at Vermont Genetic Network (VGN) Proteomics Facility. Protein concentrations were determined by Bradford assay.

Deacetylation Assay

The K_m and k_cat of SIRT6 were measured for both NAD⁺ and synthetic peptide substrate. A typical reaction was performed in 100 mM phosphate buffer pH 7.5 in a total volume of 50 μL. For NAD⁺ parameter measurement, the reactions contained various concentrations of NAD⁺, 1 mM H3K9Ac or p53K382Ac. For synthetic peptide substrate measurement, the reactions contained various concentrations of H3K9Ac or p53K382Ac, 800 μM NAD⁺. Reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. The samples were then injected on an HPLC fitted to a Macherey-Nagel Nucleosil C18 column. Co-substrate NAD⁺, products nicotinamide (NAM) and O-acetyl-ADP-ribose (AADPR) peaks were resolved using a gradient of 0 to 20% methanol in 20 mM ammonium acetate. Chromatograms were analyzed at 260 nm. Reactions are quantified by integrating areas of peaks corresponding to NAD⁺ and deacetylation product AADPR. Rates were plotted as a function of substrate concentration and best fits of points to the Michaelis-Menten equation were performed by Kaleidagraph®.

Nicotinamide Inhibition Assay

To determine nicotinamide inhibition, reactions were performed in 100 mM phosphate buffer pH 7.5 containing 800 μM NAD⁺, 500 μM H3K9Ac or 800 μM p53K382Ac, and various concentrations of NAM. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. The samples were then injected on an HPLC fitted to a Macherey-Nagel Nucleosil C18 column. Acetylated and deacetylated peptides were resolved using a gradient of 10% to 40% acetonitrile in 0.1% TFA. Chromatograms were analyzed at 215 nm. Reactions were quantified by integrating area of peaks corresponding to acetylated and deacetylated peptides. Rates were plotted as a function of NAM concentration, and points were fitted to the equation:

$$\mathrm{\nu }={\mathrm{\nu }}_0-{\mathrm{\nu }}_{\text{inh}}(\frac{[I]}{K_{\mathrm{i}}+[I]})$$

where ν is the rate observed for a given concentration of NAM, ν₀ is the uninhibited rate, ν_inh is the maximal inhibition, K_i is the apparent inhibition constant, and [I] is the concentration of NAM.

¹⁴C-Nicotinamide Base Exchange Assay

The reactions were carried out in 100 mM phosphate buffer pH 7.5 containing 800 μM NAD⁺, 1 mM H3K9Ac or p53K382Ac, 300,000 cpm [carbonyl-¹⁴C]-nicotinamide (¹⁴C-NAM, American Radiolabeled Chemicals Inc.), and various concentrations of NAM. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. The samples were then injected on an HPLC fitted to a Macherey-Nagel Nucleosil C18 column. NAD⁺ and NAM were resolved using a gradient of 0 to 20% methanol in 0.1% TFA. Fractions containing NAM and NAD⁺ were collected and the radioactivity determined by scintillation counting. Rates were expressed as cpm/s incorporated into NAD⁺, and converted to turnover rate (s⁻¹) after adjustment for specific radioactivitiy and enzyme concentration. Rates were plotted as a function of NAM concentration and best fits of points to the Michaelis-Menten equation were performed by Kaleidagraph®.

Sirtuin Inhibition Assay

A typical reaction contained 800 μM NAD⁺, 500 μM peptide substrate (H3K9Ac for SIRT2, SIRT3 and SIRT6, p53K382Ac for SIRT1 and SIRT5), varying concentrations of trichostatin A (TSA) 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 so that the conversion of substrate was less than 15%. 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 are quantified by integrating areas of peaks corresponding to NAD⁺ and AADPR. Rates were plotted as a function of TSA concentration, and points were fitted to the Morrison’s quadratic equation:

$$\frac{{\nu }_i}{{\nu }_0}=1-\frac{({[E]}_T+{[I]}_T+K_i^{\mathit{app}})-{\left({\left({[E]}_T+{[I]}_T+K_i^{\mathit{app}}\right)}^2-4{[E]}_T{[I]}_T\right)}^{0.5}}{2{[E]}_T}$$

Where ν_i is the inhibited turnover, ν₀ is the uninhibited turnover, [E]_T is the total concentration of sirtuin, [I]_T is the total inhibitor concentration, $K_{\mathrm{i}}^{\text{app}}$ is the apparent inhibition constant. This equation has been used to analyze the tight binding inhibition.^35,36 The affinity is determined in terms of free and bound concentrations of enzyme and inhibitor accounting for the effect of binary complex formation on the concentration of free inhibitor.

In addition to TSA, several other small molecules including sodium butyrate, valproic acid and SAHA were also tested against SIRT6. Reactions were carried out in 100 mM phosphate buffer pH 7.5 containing 800 μM NAD⁺, 500 μM H3K9Ac, various concentrations of the small molecules: 0, 0.5, 2 and 5 mM of sodium butyrate or valproic acid; 0, 0.1, 1, 10, 50, 100 and 200 μM of SAHA. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. 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 are quantified by integrating areas of peaks corresponding to NAD⁺ and AADPR.

Lineweaver-Burk Double-reciprocal Plot Analysis

Peptide titration reactions containing either 0, 10 μM, or 20 μM TSA were incubated with 800 μM NAD⁺, varying concentrations of H3K9Ac in 100 mM phosphate buffer pH 7.5. Similarly, NAD⁺ titration reactions containing either 0, 10 μM, or 20 μM TSA were incubated with 1 mM H3K9Ac, varying concentrations of NAD⁺ in 100 mM phosphate buffer pH 7.5. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. 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 are quantified by integrating areas of peaks corresponding to NAD⁺ and AADPR. Double reciprocal plots were generated using Kaleidagraph® and fit to a linear curve representative of the Lineweaver-Burk relationship.

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.

Nuclear Extract Preparation

HEK293 cells were treated with 1 μM TSA and 1 mM methyl methanesulfonate (MMS) at 37°C for 6 h. The cells were harvested and resuspended in a cold buffer containing 10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM EDTA, 1 mM DTT, 0.5 mM PMSF, 1 μM pepstatin A, 1 μM leupeptin, and 0.5% Tritonx100. The cells were resuspended by pipetting, incubated on ice for 10 min, and then centrifuged at 12,000 rpm for 10 min at 4°C. The supernatant was removed. The pellet was resuspended in a cold buffer containing 20 mM HEPES pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM DTT, 0.5 nM PMSF, 10% glycerol, 1 μM pepstatin A, and 1 μM leupeptin, and incubated on ice for 20 min with occasional vortexing. The sample was then centrifuged at 12,000 rpm for 20 min at 4°C. The supernatant was aliquoted and stored at −80°C. Protein concentration in the nuclear extract was determined by Bradford assay.

Western Blot

Nuclear extract (15 μg) was incubated with 1 mM NAD⁺, varying concentrations of TSA, and 20 μM SIRT6 at 37°C for 2 h. The sample was then resolved on a 12% SDS-PAGE gel and transferred to Immobilon PVDF transfer membrane (Millipore). The blot was blocked with 5% nonfat milk, probed with primary antibody targeting H3K9Ac (Sigma-Aldrich), H3 (Abcam), p53K382Ac (Cell Signaling), p53K381Ac (Abcam), p53K373Ac (Abcam), p53K372Ac (Assay Biotechnology), p53K120Ac (Abcam) or p53 (Abcam), washed with PBST, followed by incubation with anti-rabbit or anti-mouse HRP conjugated secondary antibody. The signal was then be detected by SuperSignal West Pico Chemiluminescent substrate (Thermo Fisher Scientific).

RESULTS AND DISCUSSION

SIRT6 deacetylates H3K9Ac and p53K382Ac in vitro

We have expressed and purified full length human SIRT6 with an N-terminal His-tag. This construct indeed has the ability to deacetylate a short peptide sequence which is identical to the N-terminus of histone H3 bearing the acetylated lysine 9 mark (H3K9Ac, Fig. 1B), consistent with previous report.^14,37

Figure 1.

SIRT6 deacetylates H3K9Ac and p53K382Ac in vitro. A. SIRT6 catalyzed NAD⁺-dependent deacetylation. B. HPLC chromatograms showing the deacetylation of H3K9Ac peptide by SIRT1 and SIRT6. C. HPLC chromatograms showing the deacetylation of p53K382Ac peptide by SIRT1 and SIRT6. D. HPLC chromatograms showing the production of NAM and AADPR in SIRT6 catalyzed deacetylation of H3K9Ac (a) and p53K382Ac (b). Trace c is the control experiment with no enzyme present.

Interestingly, this construct was also able to deacetylate a synthetic peptide substrate encompassing acetylated lysine 382 of p53 (p53K382Ac, Fig. 1C). SIRT1 was used as a positive control because p53 is a known physiological substrate of SIRT1. HPLC-based assay was employed to analyze the enzymatic reactions. Incubation of SIRT6 with p53K382Ac and NAD⁺ led to the formation of a new peak with similar retention time as deacetylated p53K382 comparing to SIRT1 catalyzed reaction (Fig. 1C). The identity of the peak was further confirmed by MS analysis. Moreover, when SIRT6 catalyzed reaction was examined at 260 nm, a fork-like peak was observed on the chromatogram (Fig. 1D). Sirtuins consume stoichiometric amount of NAD⁺ to remove acetyl group from lysine residues, and to produce nicotinamide (NAM) and O-acetyl-ADP-ribose (AADPR, Fig. 1A).^38,39 In aqueous solution 2′-AADPR and 3′-AADPR equilibrate rapidly, resulting in the signature fork-like peak.⁴⁰ All of these data support the notion that SIRT6 could deacetylate p53 at lysine 382 in vitro. Notably, SIRT6 demonstrated specific deacetylation of p53K382, but had no activity on the other acetylated p53 peptides tested (Fig. S1). Acetylated lysine 120, 372, 373, and 381 have been selected for their critical roles in maintaining p53 stability, DNA binding affinity as well as transcriptional activity.⁴¹ However, none of them can be deacetylated by SIRT6 to appreciable level (Fig. S1).

The kinetic constants for SIRT6 mediated deacetylation under steady state conditions at pH 7.5 have been determined (Table 1 and Fig. 2). SIRT6 demonstrated low catalytic efficiency towards both peptide substrates, which stemmed primarily from the low binding affinity. This low activity corroborates well with the recent discovery that SIRT6 harbors other enzymatic activities in addition to the deacetylase activity.^22,23

Table 1.

Recombinant SIRT6 steady state parameters

	deacetylation				base exchange
Substrate	Km (μM)	kcat (10−3s−1)	kcat/Km (s−1M−1)	Ki(NAM) (mM)	Km (mM)	kcat (10−4s−1)
H3K9Ac	1665 ± 115	1.8 ± 0.2	1.1	2.25 ± 0.6	2.46 ± 0.3	2.1 ± 0.1
p53K382Ac	1598 ± 84	1.4 ± 0.6	0.9	3.19 ± 0.4	1.82 ± 0.5	2.8 ± 0.3

Figure 2.

Steady-state kinetic analysis of SIRT6 catalyzed deacetylation. Peptide saturation curves for H3K9Ac (circle) and p53K382Ac (square). Reactions containing 800 μM NAD⁺, and various concentrations of H3K9Ac or p53K382Ac in 100 mM phosphate buffer pH7.5 were initiated by the addition of 10 μM of SIRT6. The reactions were incubated at 37°C for 2 h and quenched by 8 μL of 10% TFA. Rates were determined as described in “Methods and Material”. Rates were plotted as a function of peptide concentration and fit to the Michaelis-Menten equation using Kaleidagraph®. Error bars represent S.D. of at least three replicates.

SIRT6 deacetylates full length histone H3 and p53

We then investigated the deacetylase activity of SIRT6 on physiological substrates, namely full length histone H3 and p53. HEK293 cells were cultured in the presence of methyl methanesulfonate (MMS), an alkylating agent that causes double-strand DNA breaks. MMS treatment is known to cause hyperacetylation of nuclear and cytoplasmic proteins.⁴² The nuclear extract isolated from these cells was incubated with recombinant SIRT6 and NAD⁺. The acetylation levels of H3K9 and p53K382 were probed using specific antibodies (Fig. 3A and 3B). SIRT6 demonstrated robust deacetylation activity towards H3K9 in full length H3 (Fig. 3A). The deacetylation of p53K382 was also detected, although to a lesser extent (Fig. 3B). Similar to the in vitro testing, SIRT6 specifically deacetylated lysine 382, but not the other p53 lysine residues probed (Fig. S2). Together, our results suggest that SIRT6 is a NAD⁺-dependent deacetylase targeting p53 lysine 382. Depletion of acetylated p53K382 has been associated with the reduction of sequence-specific DNA-binding,⁴³ with the suppression of p53-mediated apoptosis following DNA damage insults,³⁰ and with ubiquitination-triggered p53 degradation.^44,45 In addition, hyperacetylation of p53K382 was observed upon FK866 treatment, most likely via sirtuin inhibition.⁴⁶ FK866 significantly decreased intracellular NAD⁺ level through inhibition of NAMPT. Loss of NAD⁺ led to the reduction of sirtuin deacetylation activity. SIRT6 may sense the cellular energy fluctuation through NAD⁺ level changes and elicit profound responses by deacetylating downstream targets such as p53K382. Further investigation of the physiological relevance of deacetylation of p53K382 by SIRT6 is in progress.

Figure 3.

SIRT6 deacetylates full length H3 and p53. A. Western blots showing SIRT6 deacetylates full length histone H3. B. Western blots showing SIRT6 deacetylates full length p53. The gel images shown represent two replicates with the same condition. C. Quantification of A and B. The data represents average of as least three independent experiments ± S.D. Statistical significance was determined by a Student’s t-test: *p < 0.01 vs nuclear extract alone.

Nicotinamide partially inhibits SIRT6 deacetylase activity

The importance of SIRT6 in various cellular processes has magnified the need to dissect the molecular mechanism underlying its biological functions. It has been well established that sirtuin catalyzed reaction goes through an ADP-ribosyl-peptidyl imidate intermediate.^38,47 The imidate can either proceed forward to achieve deacetylation, or to combine with NAM to regenerate NAD⁺ via a process called “base exchange”. Depletion of the imidate intermediate through base exchange could inhibit deacetylation.⁴⁷ When NAM was titrated into SIRT6 catalyzed reactions, partial inhibition of deacetylation was detected (Fig. 4A). K_i was determined to be 2.25 and 3.19 mM for H3K9Ac and p53K382Ac, respectively, suggesting that SIRT6 deacetylase activity is unlikely to be influenced by physiological NAM concentration which is normally in the mid-micromolar range.^48,49 The K_i values for SIRT6 were at least 10 times higher than those for SIRT1, SIRT2 and SIRT3.^50,51,52 Archeaglobus fulgidus Sir2 (Af2Sir2) also demonstrated incomplete inhibition of deacetylation by NAM.⁴⁷ The K_i value for Af2Sir2 was in the low micromolar range (26 μM) with no additional inhibition beyond mM concentration. The lack of inhibition has been explained as a result of a slow base exchange incapable of competing with the fast forward deacetylation.⁴⁷ However, for SIRT6, inhibition only became apparent after NAM concentration reached mM range (Fig. 4A). And this inhibition was not substrate specific since the K_i values for H3K9Ac and p53K382Ac were comparable. Similar insensitivity to NAM inhibition has been observed for SIRT5.⁵³

Figure 4.

Nicotinamide base exchange catalyzed by SIRT6. A. Inhibition of the deacetylation by increasing concentrations of nicotinamide. The reactions were performed in 100 mM phosphate buffer pH 7.5 containing 800 μM NAD⁺, 500 μM H3K9Ac (circle) or 800 μM p53K382Ac (square), and various concentrations of NAM. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. Rates were determined as described in “Methods and Material” and were plotted as a function of NAM concentration. The points were fitted to the equation: $\mathrm{\nu }={\mathrm{\nu }}_0-{\mathrm{\nu }}_{\text{inh}}(\frac{[I]}{K_{\mathrm{i}}+[I]})$ using Kaleidagraph®. B. Kinetics of SIRT6 catalyzed base exchange of [carbonyl-¹⁴C]-nicotinamide into unlabeled NAD⁺. The reactions were carried out in 100 mM phosphate buffer pH 7.5 containing 800 μM NAD⁺, 1 mM H3K9Ac or p53K382Ac, 300,000 cpm [carbonyl-¹⁴C]-nicotinamide, and various concentrations of NAM. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. Rates were determined as described in “Methods and Material” and were plotted as a function of NAM concentration. The pointes were fitted to the Michaelis-Menten equation using Kaleidagraph®. Error bars represent S.D. of at least three replicates.

The base exchange reaction was further characterized using carbonyl-¹⁴C labeled NAM as described before.^47,54 The rate of incorporation of exogenous ¹⁴C-NAM into NAD⁺ in the presence of peptide substrate and SIRT6 was measured. As expected, the formation of ¹⁴C-NAD⁺ increased with increasing concentration of NAM (Fig. 4B). The K_m value was determined to be 2.46 mM and 1.82 mM for H3K9Ac and p53K382Ac, respectively. The corresponding k_cat values were 2.1 × 10⁻⁴ s⁻¹ (H3K9Ac) and 2.8 × 10⁻⁴ s⁻¹ (p53K382Ac), only a fraction of the k_cat values of deacetylation (Table 1). Thus the insensitivity of SIRT6 deacetylation to NAM inhibition can be attributed to the kinetically incompetent base exchange, and appreciable inhibition can only be achieved at non-physiological NAM concentration.

TSA is a competitive inhibitor of SIRT6

Small molecule sirtuin inhibitors have significantly empowered the study of sirtuin biology,^55,56 and ultimately will eventuate in the development of sirtuin-targeted therapeutics. However, inhibitors targeting SIRT6 have been heavily investigated with very little success.^27,28,29 This has prompted us to direct our effort at the discovery of SIRT6 inhibitors. TSA is a potent and selective Class I/II HDAC inhibitor with the IC₅₀ in the low nanomolar range (Table 2).^57,58,59 It has been widely used in preclinical study to interrogate the function of histone acetylation in regulating gene expression. TSA has also been suggested as a potential therapeutic for the treatment of cancer, cardiovascular diseases and neurodegenerative disorders. However, due to its undesired pharmacological profile, TSA has not been used in clinical trials. In an effort to develop SIRT6-selective inhibitor, TSA was assessed for its effect on SIRT6 catalyzed deacetylation using the HPLC assay described in “Methods and Materials”. With H3K9Ac as the substrate, TSA was able to significantly inhibit the deacetylase activity in a dose-dependent manner as evidenced by the reduction of deacetylated H3K9 as well as AADPR. The K_i for both H3K9Ac and p53K382Ac was measured to be 2.02 μM and 4.62 μM, respectively (Table 2 and Fig. 5A). This inhibition is unique to SIRT6 since no appreciable activity reduction was detected for SIRT1, SIRT2, SIRT3 and SIRT5 at up to 50 μM TSA (Fig. S3).

Table 2.

Known Class I and II HDAC inhibitors

Inhibitor	Classification	Selectivity	IC50 (μM)					Ki (μM)	Reference
			HDAC1	HDAC3	HDAC4	HDAC6	HDAC8	SIRT6
TSA	hydroxyamate	Class I/II	0.0015	0.0006	0.038	0.0086	0.49	2.02 (H3K9Ac) 4.62 (p53K382Ac)	57,58
SAHA	hydroxyamate	Class I/II	0.0137	0.106	0.282	0.002	0.0068	> 200	57, 69,70
Sodium butyrate	aliphatic fatty acid	pan HDAC	175				350	> 5,000	71,69
Valproic acid	aliphatic fatty acid	Class I/IIa	171	1,000	1,500	>20,000	756	> 5,000	71,69

Figure 5.

TSA inhibits SIRT6 catalyzed deacetylation of synthetic peptides. A. Determination of inhibition constant K_i of TSA. Reactions containing 800 μM NAD⁺, 500 μM or 5 mM H3K9Ac, varying concentrations of TSA in 100 mM phosphate buffer pH 7.5 were initiated by the addition of 10 μM of SIRT6, and were incubated at 37°C before being quenched by 8 μL of 10% TFA. Rates were determined as described in “Methods and Material” and plotted as a function of TSA concentration. Points were fitted to the Morrison’s quadratic equation to get the inhibition constant. B and C are the double-reciprocal inhibition plots. B. TSA exhibits competitive inhibition towards H3K9Ac in SIRT6 catalyzed reaction. Reactions containing either 0, 10 μM, or 20 μM TSA were incubated with 800 μM NAD⁺, varying concentrations of H3K9Ac in 100 mM phosphate buffer pH 7.5. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. Rates were determined as described in “Methods and Material”. C. TSA exhibits non-competitive inhibition towards NAD⁺ in SIRT6 catalyzed reactions. Reactions containing either 0, 10 μM, or 20 μM TSA were incubated with 1 mM H3K9Ac, varying concentrations of NAD⁺ in 100 mM phosphate buffer pH 7.5. The reactions were initiated by the addition of 10 μM of SIRT6 and were incubated at 37°C for 2 h before being quenched by 8 μL of 10% TFA. Rates were determined as described in “Methods and Material”. Double reciprocal plots were generated using Kaleidagraph® and fit to a linear curve representative of the Lineweaver-Burk relationship. Error bars represent S.D. of three replicates.

Class I and II HDACs utilize an active-site Zn²⁺ to coordinate to acetyl group of lysine substrate,⁶⁰ and to facilitate the subsequent hydrolysis of acetyllysine. TSA was thought to use the hydroxamic acid “warhead” to sequester the active-site zinc, resulting in mechanism-based inhibition. It was surprising to see that TSA was able to inhibit Class III HDAC, namely sirtuins, since they employ a completely different deacetylation mechanism. Sirtuin binds to both co-substrate NAD⁺ and acetylated peptide substrate to initiate catalysis. An imidate intermediate is formed following the cleavage of N-glycosic bond in NAD⁺. The subsequent collapse of the imidate intermediate leads to the formation of deacetylated lysine product, at the same time NAD⁺ is converted to AADPR. Sirtuins conserve a catalytic core domain which contains approximately 270 amino acids. Accumulating structural evidence has shown that sirtuin consists of two distinct domains: a larger Rossmann fold which is a classic pyridine dinucleotides binding fold, and a smaller domain composed of a zinc-binding module and a flexible helical module.^61,37 Between the Rossmann fold domain and the smaller domain is a cleft where acetyllysine peptide and NAD⁺ cofactor bind. The acetyllysine inserts into a hydrophobic tunnel which is immediate adjacent to the NAD⁺ binding site. The presence of the zinc-binding module in sirtuin is thought to maintain structural integrity, but not to participate in catalysis.⁶¹ The initial hypothesis was that TSA may have targeted zinc in the small domain. However, the selective inhibition of SIRT6, but not other human sirtuins suggests an alternative mechanism. Supplementation of zinc to TSA bound-SIRT6 led to negligible reduction of TSA-induced SIRT6 inhibition (Fig. S4), further confirming that hydroxamic acid moiety in TSA serves more than just a metal chelator. Furthermore, several other Class I/II HDAC inhibitors including another hydroxamate, SAHA, and two short chain aliphatic fatty acids, valproic acid and sodium butyrate, were profiled against SIRT6 in a dose-ranging format. None of these compounds demonstrated noticeable inhibitory effect (Fig. S5). It appeared that TSA selectively inhibits SIRT6 among the human sirtuins. And the regulatory mechanism of this inhibition is different from those for the Class I and II HDACs.

The type of inhibition was then characterized. The initial rate of AADPR formation was measured at saturating NAD⁺ concentration with varying concentrations of H3K9Ac and TSA. For each TSA concentration, 1/rate was plotted as a function of 1/[H3K9Ac] (Fig. 5B). Double-reciprocal plot demonstrated that TSA is competitive with the acetylated peptide substrate as revealed by a series of lines that intersected at the y-axis. Similarly, the initial rate was assessed at fixed H3K9Ac concentration but various NAD⁺ and TSA levels. The lines intersected at the x-axis indicating that TSA is non-competitive with respect to NAD⁺ (Fig. 5C).

Base-exchange in the presence of TSA was also assayed using radiolabeled NAM (Fig. S6). SIRT6 was incubated with H3K9Ac peptide, NAD⁺, ¹⁴C-NAM, with or without TSA. The formation of ¹⁴C-NAD⁺ was quantified by scintillation counting. TSA competed with peptide substrate to impede the formation of imidate intermediate, and ultimately resulted in the reduction of base-exchange. This inhibition had very little effect on the K_m of base-exchange, suggesting that TSA and NAM may occupy distinct binding sites.

TSA inhibits SIRT6 deacetylation of full length H3 and p53

The ability of TSA to influence SIRT6 catalyzed deacetylation of full length H3 and p53 was investigated. A combination of nuclear extract from HEK293 cells, SIRT6 and NAD⁺ was incubated with or without TSA. Western blots analysis revealed that the acetylation levels of both H3K9 and p53K382 were elevated with increasing TSA concentration (Fig. 6A). This induction was noticeable even at the lowest concentration (1 μM). TSA treatment was more effective in restoring the acetylation level of p53K382 than that of H3K9 (Fig. 6B).

Figure 6.

TSA inhibits SIRT6 catalyzed deacetylation of full length H3 and p53. Representative western blot (A) and quantification analysis (B) showing that TSA inhibits the deacetylation of H3K9 and p53K382 in HEK293 cell nuclear extract in a dose-dependent manner. Representative western blot (C) and quantification analysis (D) showing that NAM only marginally inhibits SIRT6 catalyzed deacetylation of H3K9 and p53K382 in HEK293 cell nuclear extract. The quantification data represents average of three independent experiments ± S.D. Statistical significance was determined by a Student’s t-test: *p < 0.01 vs nuclear extract alone; ^#p < 0.01 vs SIRT6 and NAD⁺ treatment; ^‡p < 0.05 vs SIRT6 and NAD⁺ treatment.

NAM inhibition was probed using physiological substrates as well. At millimolar concentrations, NAM was capable of impeding SIRT6 mediated deacetylation (Fig. 6C). This inhibition seemed to plateau at high concentrations (~5 mM, Fig. 6D), in agreement with the results using synthetic peptides.

CONCLUSIONS

SIRT6 is a chromatin-associated epigenetic modification enzyme that has been implicated in the regulation of gene silencing, genome stability, DNA repair as well as glucose homeostasis.⁶² Despite the intense pursuit of SIRT6 as a novel therapeutic target, little is known about its cellular targets and biological functions. This is partly due to the fact that SIRT6 demonstrates weak deacetylase activity in vitro, making the identification to its substrates extremely challenging. In the current study, p53 lysine 382 has been identified as a substrate of human SIRT6. Recombinant enzyme was able to deacetylate K382 in short peptide sequence as well as in full length p53. The removal of acetyl group is very selective for lysine 382 as SIRT6 failed to demonstrate appreciable deacetylation on the other lysine residues tested. Kinetic analysis revealed low deacetylation efficiency, similar to that of the known SIRT6 substrate, H3K9. This low catalytic activity has been attributed to a splayed configuration between the Rossmann fold domain and the metal binding domain.³⁷ Posttranslational modifications such as ubiquitination, methylation and acetylation of p53 have been suggested to dictate its tumor suppressor function.^63,41 Acetylated lysine 382 has been identified as the endogenous substrate of HDAC1⁶⁴ and SIRT1.^30,31 Deacetylation at this specific lysine residue can cause ubiquitin-dependent p53 degradation,⁴⁴ and ultimately reduces its transcriptional activity.^30,31 Another speculative function of p53 acetylation is to recruit deacetylases for targeted gene repression.³¹ The regulation of p53K382 acetylation remains largely elusive. Other protein deacetylases may also control its acetylation level. Our finding suggested the potential role of SIRT6 in cell apoptosis and stress resistance through direct deacetylating p53K382. Additional evidence is needed to establish the biological significance of this modification.

To better elucidate the biochemical and biological function of SIRT6, small molecule chemical probes were interrogated. NAM is a well-established negative physiological regulator of sirtuins.⁴⁷ It competes with deacetylation for the same imidate intermediate. NAM was found to be a weak inhibitor of SIRT6 with K_i values in the millimolar range. This can be explained as a result of low affinity of NAM to the enzyme supported by the high K_m values acquired in the base exchange studies.

Strikingly, in a search of SIRT6 regulator with novel structural framework and isoform specificity, TSA has been identified as a potent and selective (among sirtuin family) SIRT6 inhibitor with K_i values in the low micromolar range. This inhibition was not provided by several other Class I/II HDAC inhibitors. Previous studies have suggested that SIRT6 mRNA and protein expression levels can be regulated by TSA treatment.^65,66,67 However, our data indicated, for the first time, that TSA directly acts on SIRT6 to control its enzymatic activity. In-depth kinetic analysis indicated that TSA serves as a competitive inhibitor rather than a mere metal-chelator. The ultimate goal of our study is to develop SIRT6-selective inhibitors. Knowledge on the binding site and binding mode of TSA will provide blueprint for the design of more potent inhibitors. Currently we are taking a two-pronged approach. On one hand, TSA will be co-crystalized with SIRT6. It will provide critical information on the binding site of TSA, and suggest possible structure optimization strategies. On the other hand, several TSA analogs bearing photoaffinity groups have been synthesized. Photoaffinity labeling of proteins/enzymes with small molecule probes represents a unique strategy to study ligand-protein or substrate-enzyme interactions.⁶⁸ Labeling of SIRT6 with TSA analogs, combined with MS analysis, may pinpoint the amino acid residues that are responsible for making direct contact with TSA. Information from these studies will guide the design and synthesis of the next generation of inhibitors. All of these can be expected from our group in the near future.

ABBREVIATIONS

NAD⁺

nicotinamide adenine dinucleotide

NAM

nicotinamide

AADPR

O-acetyl-ADP-ribose

TSA

trichostatin A

HDAC

histone deacetylase

TFA

trifluoroacetic acid

MMS

methyl methanesulfonate