Multiple Regulatory Layers of SREBP1/2 by SIRT6 - Cell Press

Cell Reports

Report Multiple Regulatory Layers of SREBP1/2 by SIRT6 Sivan Elhanati,1 Yariv Kanfi,1 Alexander Varvak,1 Asael Roichman,1 Ilana Carmel-Gross,1 Shaul Barth,1 Gilad Gibor,1 and Haim Y. Cohen1,* 1The Mina & Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan 52900, Israel *Correspondence: [email protected] http://dx.doi.org/10.1016/j.celrep.2013.08.006 This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial-No Derivative Works License, which permits non-commercial use, distribution, and reproduction in any medium, provided the original author and source are credited.

SUMMARY

The NAD+-dependent protein deacetylase SIRT6 regulates genome stability, cancer, and lifespan. Mice overexpressing SIRT6 (MOSES) have lower low-density lipoprotein cholesterol levels and are protected against the physiological damage of obesity. Here, we examined the role of SIRT6 in cholesterol regulation via the lipogenic transcription factors SREBP1 and SREBP2, and AMP-activated protein kinase (AMPK). We show that SIRT6 represses SREBP1 and SREBP2 by at least three mechanisms. First, SIRT6 represses the transcription levels of SREBP1/SREBP2 and that of their target genes. Second, SIRT6 inhibits the cleavage of SREBP1/SREBP2 into their active forms. Third, SIRT6 activates AMPK by increasing the AMP/ATP ratio, which promotes phosphorylation and inhibition of SREBP1 by AMPK. Reciprocally, the expression of miR33a and miR33b from the introns of SREBP2 and SREBP1, respectively, represses SIRT6 levels. Together, these findings explain the mechanism underlying the improved cholesterol homeostasis in MOSES mice, revealing a relationship between fat metabolism and longevity. INTRODUCTION Sirtuins are homologs of the yeast NAD+-dependent SIR2 deacetylase, which was shown to regulate yeast lifespan. The seven mammalian sirtuins, SIRT1–7, were implicated in the regulation of various aging-related pathways such as cellular senescence, inflammation, and genome stability. However, to date, only SIRT6 was shown to directly regulate mammalian lifespan (Kanfi et al., 2012). SIRT6 possesses histone H3 K9 and K56 deacetylase activity (Michishita et al., 2008, 2009; Yang et al., 2009). On chromatin, SIRT6 functions to attenuate NF-kB signaling by deacetylating H3K9 on NF-kB target gene promoters (Kawahara et al., 2009). Similarly, SIRT6 also represses HIF-1a-regulated genes by deacetylating their promoters (Zhong et al., 2010). As a result, SIRT6 depletion enhances glycolysis due to an increase in the

expression levels of glycolytic genes. Many cancer cells produce energy by increasing aerobic glycolysis instead of oxidative metabolism, a phenomenon known as the ‘‘Warburg effect’’ (Warburg, 1956). Thus, SIRT6 also functions as a tumor suppressor that regulates aerobic glycolysis in cancer cells (Sebastiań et al., 2012). Interestingly, SIRT6 regulates other cancer-related pathways, including genome stability, by modulating DNA double-strand break repair, base excision repair (BER), and telomere maintenance (Kawahara et al., 2009; Mostoslavsky et al., 2006). Mice genetically deficient in SIRT6 display a set of prematureaging-like phenotypes, such as lymphopenia, loss of subcutaneous fat, and lordokyphosis and die at about 4 weeks of age (Mostoslavsky et al., 2006). The mice also exhibit hypoglycemia as a result of a significant increase in glucose uptake due to higher expression levels of the glucose transporter, GLUT1. Previous results have shown SIRT6 to be involved in the dietary restriction response that is known to extend mouse lifespan (Kanfi et al., 2008). In addition, when fed a high-fat diet (HFD), transgenic mice overexpressing SIRT6 (MOSES), in comparison to their wild-type (WT) littermates, accumulate significantly less visceral fat, low-density lipoprotein (LDL) cholesterol, and triglycerides (Kanfi et al., 2010). The effect of SIRT6 on cholesterol levels is not limited to HFD. In comparison to their WT littermates, MOSES mice have significantly lower cholesterol levels even when fed a regular chow diet. Thus, the mechanism underlying the regulation of cholesterol homeostasis by SIRT6 may shed further light on its metabolic effects. Sterol regulatory element binding proteins (SREBPs) are lipogenic transcription factors that are regulated by cholesterol, insulin, and glucose (Horton et al., 2002). The mammalian genome contains two distinct SREBP genes: SREBP1 and SREBP2. SREBP1 mainly regulates lipogenic processes by activating genes involved in fatty acid and triglyceride biosynthesis, whereas SREBP2 mostly activates genes involved in cholesterol synthesis. The SREBP1 gene produces two different isoforms, which differ in their first exon, SREBP1a and SREBP1c, owing to the use of different transcription start sites on the SREBP1 gene (Miserez et al., 1997). SREBPs become active upon proteolytic processing of the inactivate precursor in the endoplasmic reticulum (ER), which releases a transcriptionally active N-terminal basic helix-loop-helix (bHLH) zip domain. A complex comprising S1P and S2P proteases and SCAP cleaves SREBPs into their active forms. The mature form translocates to the nucleus and promotes a lipogenic program in the liver. SREBP activates itself and its target genes by binding to sterol regulatory

Cell Reports 4, 905–912, September 12, 2013 ª2013 The Authors 905

elements (SREs) located at the various promoters. Therefore, dysregulation of SREBPs was shown to be involved in type-2 diabetes, dyslipidemia, and hepatic steatosis (Raghow et al., 2008). The heterotrimeric serine/threonine kinase, AMPK (AMPactivated protein kinase), is one of the major energy sensors in eukaryotic cells (Kahn et al., 2005). AMPK is ubiquitously expressed and functions as an intracellular fuel sensor, thereby maintaining energy balance. AMPK is highly sensitive to changes in AMP levels. Therefore, a slight increase in the AMP/ATP ratio results in AMPK’s activation via the phosphorylation of Thr172 by its kinases, LKB and CaMKK (Corton et al., 1994; Hardie et al., 1999). Recently, it was shown that the ADP/ATP ratio also contributes to the regulation of AMPK (Oakhill et al., 2011). Once activated, AMPK phosphorylates a wide variety of downstream targets, resulting in the overall outcome of activating ATP-producing processes, while extinguishing ATP-consuming ones (Merrill et al., 1997). The net activation of AMPK results in hepatic fatty acid oxidation, and the inhibition of cholesterol and triglyceride synthesis, and lipogenesis. SREBP1 activity was recently shown to be negatively regulated by AMPK. Phosphorylation of SREBP1 on Ser372 by AMPK inactivates SREBP1 by suppressing its cleavage and nuclear translocation and attenuates hepatic steatosis and atherosclerosis in mice (Li et al., 2011). In this study, we examine the role of SIRT6 in cholesterol regulation via SREBP1/SREBP2 and AMPK. We show that SIRT6 represses SREBP1 and SREBP2 on at least three levels. First, SIRT6 represses the transcription of SREBP1 and SREBP2 and their target genes. Second, SIRT6 inhibits the cleavage of SREBP1 and SREBP2 into their active forms. Finally, SIRT6 activates AMPK by increasing the AMP/ATP ratio, promoting the phosphorylation and inhibition of SREBP1. Together, these findings explain the mechanism underlying the improved cholesterol homeostasis in MOSES mice, revealing a connection between energy homeostasis and longevity. RESULTS Previously, we showed that mice overexpressing SIRT6 have lower LDL cholesterol levels in comparison to their WT littermates (Kanfi et al., 2010). To examine if this effect is systemic or cell autonomous, we followed the levels of cholesterol and triglycerides in hepatoma HepG2 cells transfected with an empty vector or a SIRT6 expression plasmid. As seen in Figure 1A, cholesterol (left panel) and triglyceride (right panel) levels were significantly lower in cells overexpressing SIRT6. Conversely, knockdown of SIRT6 by small interfering RNA (siRNA) resulted in a significant increase in cholesterol (left panel) and triglyceride (right panel) levels (Figure 1B). To further examine the role of SIRT6 in regulating fat homeostasis, the expression levels of Figure 1. SIRT6 Overexpression Reduces Cholesterol and Triglyceride Levels and Expression of Lipogenic Genes in Hepatocytes Cholesterol (left panel) or triglyceride (right panel) levels in (A) HepG2 cells transfected with SIRT6 or empty vector, and in (B) HepG2 cells transfected with control siRNA (si-control) or siRNA against SIRT6 (si-SIRT6). Representative western blots demonstrating SIRT6 levels are shown on the right. Expression profile (qRT-PCR) of fatty acid and triglyceride biosynthesis genes in (C) HepG2 cells overexpressing SIRT6, relative to cells transfected with

empty vector (WT), and in (D) HepG2 cells knocked down for SIRT6, relative to cells with si-control (WT). A ChIP experiment followed by qRT-PCR measuring (E) SIRT6 binding or (F) histone H3K56Ac levels at the promoter regions of SREBP1, SREBP2, and their target genes, in HepG2 cells overexpressing empty vector (WT) or SIRT6. An anti-H3 Ab was used as a positive control (data not shown); anti-GFP was used as a negative control (N.C.). Data represent means ± SEM, *p % 0.05. See also Figure S1.

906 Cell Reports 4, 905–912, September 12, 2013 ª2013 The Authors

Figure 2. Effect of SIRT6 Overexpression on Lipogenic Gene Transcription Mediated by AMPK (A) Representative western blot analysis of SREBP1/SREBP2, AMPK, p-AMPK (Thr172), ACC, p-ACC (Ser 79) and SIRT6 in HepG2 cells overexpressing SIRT6 or empty vector. P and N designate the precursor (125 kDa) and nuclear (65 kDa) forms, respectively, of SREBP1/2. Densitometric quantification of p-AMPK, p-ACC, and pSREBP1 from three different blots is shown below. (B) qRT-PCR analysis of SCAP, S1P, and S2P under SIRT6 overexpression (gray bars), relative to vector alone (black bars). (C) ChIP experiment followed by qRT-PCR measuring SIRT6 binding at the promoter regions of SCAP, S1P, and S2P in HepG2 cells overexpressing empty vector (black bars) or SIRT6 (gray bars). (D) Western blot analysis of S1P, S2P, and SCAP in HepG2 cells overexpressing empty vector or SIRT6. (E) AMP/ATP ratio in HepG2 cells transfected with SIRT6 or empty vector. Results shown are the means ± SEM from five independent measurements and were normalized to 1 for the control cells. (F) Western blot analysis of pLKB, LKB, and CaMKKb in HepG2 cells overexpressing empty vector or SIRT6. (G) Expression levels of SREBP1/SREBP2 target genes in HepG2 cells transfected with SIRT6 together with scrambled RNA (gray bars) or siAMPK (dark gray bars), relative to vector alone (black bars). The efficacy of the siRNA is shown above. In all blots, b-actin served as a loading control. Data represent means ± SEM, *p % 0.05, **p % 0.01. See also Figures S2 and S3.

genes involved in fatty acid and triglyceride biosynthesis were examined in WT and SIRT6-overexpressing HepG2 cells. Quantitative real-time PCR (qRT-PCR) of the mRNA levels of SREBP1 and SREBP2 and their target genes FASN, ACC1, ACC2, LDLR, HMGCR, HMGCS, and SCD1 showed that they were significantly lower under conditions of SIRT6 overexpression (Figure 1C). Reciprocally, knockdown of SIRT6 by siRNA resulted in a significant increase in the mRNA levels of these genes (Figure 1D). Similarly, in comparison to the WT liver, the expression levels of selected SREBP1 and SREBP2 target genes, including FASN, ACC1, ACC2, LDLR, HMGCR, and HMGCS, were significantly lower in the liver of SIRT6 transgenic mice (Figure S1).

SIRT6 negatively regulates the transcription of its target genes by binding and deacetylating its promoter on histone H3 K9 and K56. Indeed, chromatin immunoprecipitation (ChIP) demonstrated that SIRT6 binds to the promoter region of SREBP1c, SREBP2, SCD1, and HMGCS (Figure 1E). Moreover, ChIP analysis showed that, in comparison to WT cells, the levels of H3K56Ac on these promoters was significantly lower in SIRT6-overexpressing cells (Figure 1F). Thus, these results suggest that SIRT6 represses the transcription of these genes by deacetylating histone H3 K56 on these promoters. Additional layers of regulation control SREBP1 and SREBP2 activity. First, posttranslational cleavage of SREBP1 and SREBP2 liberates a transcriptionally active fragment of these proteins, which translocates to the nucleus (Brown and Goldstein, 1997). In addition, SREBP1 is negatively regulated upon phosphorylation by AMPK. Therefore, we determined whether SIRT6 affects these modes of SREBP1/SREBP2 regulation. As seen in Figure 2A, overexpression of SIRT6 in HepG2 cells


A

C

Figure 3. Reciprocal Regulation between SIRT6 and miR-33

B

D

blocked the cleavage of SREBP1 and SREBP2, reducing the formation of the nuclear active forms. In addition, SIRT6 overexpression resulted in a significant reduction in the transcript levels of SCAP, S1P, and S2P, three members of the protein complex that cleaves SREBP1 and SREBP2 (Figure 2B). In support of this notion, ChIP analysis demonstrated that SIRT6 bound to the promoter region of SCAP, S1P, and S2P (Figure 2C). Similarly, SIRT6 overexpression resulted in a decrease in SCAP, S1P, and S2P protein levels (Figure 2D). Thus, SIRT6 overexpression blocks the processing and activity of SREBP1 and SREBP2. To further elucidate AMPK’s regulation of SREBP, the levels of phosphorylated Ser372 of SREBP1c, a known AMPK inhibition target, were examined. As seen in Figure 2A, SREBP1c phosphorylation was significantly elevated in cells overexpressing SIRT6. The increase in phosphorylated SREBP1c under SIRT6 overexpression prompted us to follow the levels of AMPK activity in HepG2 cells overexpressing SIRT6 versus an empty vector. In agreement with the increase in phosphorylated SREBP1c, SIRT6 overexpression resulted in a significant increase in phosphorylated AMPK (Thr172) and in its phosphorylated substrate, p-ACC (Figure 2A). qRT-PCR demonstrated that SIRT6 overexpression in vivo results in a significant increase in the transcription levels of AMPK in the liver (Figure S2). An increased level of AMPK is known to promote fatty acid beta oxidation (Viollet et al., 2009). Indeed, the levels of liver HADHB, CPT, and CROT, rate-limiting enzymes in oxidation of beta free fatty acids, were significantly increased in parallel to AMPK (Figure S2). An increased AMP/ATP ratio activates AMPK. Indeed, in comparison to control cells, the AMP/ATP ratio was 2-fold higher in SIRT6-overexpressing cells (Figure 2E). AMPK is considered ultrasensitive to ATP depletion and, consequently, to AMP elevation. Therefore, the observed 2-fold increase in AMP/ATP ratio is sufficient to activate AMPK (Hardie and Hawley, 2001). Similarly,

(A) Target sequence for miR-33a and miR-33b in SIRT6 30 UTR. (B and C) SIRT6 (B) and SREBP1/SREBP2 (C) protein levels following overexpression or downregulation of miR-33a. b-actin bands served as a loading control. SC, scrambled sequence. (D) qRT-PCR expression profile of miR-33a and miR-33b in HepG2 cells overexpressing SIRT6. (E) qRT-PCR analysis of miR-33 expression in liver from WT and SIRT6 transgenic (TG) male mice (n = 8 per group). R.I., relative intensity. Data represent the mean ± SEM. *p % 0.05.

E

SIRT6 overexpression also resulted in an increase in ADP/ATP ratio, which can also activate AMPK (Figure S3) (Oakhill et al., 2011). Activation of LKB (Shaw et al., 2005; Woods et al., 2003) or increased levels of CaMKKb (Hawley et al., 2005; Woods et al., 2005) were shown to activate AMPK, as well. As seen in Figure 2F, SIRT6 overexpression resulted in an increase in the active form of LKB (pLKB Ser428) and no change in CaMKKb levels. Together, these findings suggest that the reduction in the expression of lipogenic gene transcription due to SIRT6 overexpression is mediated, at least partially, via AMPK. To examine this possibility, we followed the effect of SIRT6 overexpression on the levels of lipogenic genes in the presence or absence of AMPK. SIRT6 overexpression significantly reduced the RNA levels of FASN, ACC1, ACC2, LDLR, HMGCR, HMGCS, SCD1, and SREBP1c in HepG2 cells. However, reduction in AMPK expression by siRNA, which targets both AMPK isoforms, restored the RNA levels of these genes (Figure 2G). Thus, the inhibitory effect of SIRT6 on expression of these genes is mediated by AMPK. Together, these findings reveal additional mechanisms by which SIRT6 inhibits the expression of lipogenic genes. SIRT6 suppresses transcriptional activation cleavage of SREBP1 and SREBP2 by reducing levels of SCAP, S1P, and S2P. In addition, by increasing the AMP/ATP ratio and stimulating AMPK phosphorylation, SIRT6 inhibits SREBP1c. MiR-33a and miR-33b are transcribed from introns of SREBP2 and SREBP1, respectively, and were shown to regulate SIRT6 levels (Figure 3A) (Da´valos et al., 2011). We therefore examined the possibility of a feedback regulation of SIRT6 by SREBP1/ SREBP2 via miR-33a and miR-33b. Given that the seed sequences of both miR’s are identical, we followed the effect of only miR-33a. HepG2 cells were transfected with miR-33a or with a scrambled control miR. As seen in Figure 3B, miR-33a overexpression resulted in a 30% reduction in SIRT6 protein levels. Conversely, transfection with antagomir against miR33a resulted in a 30% increase in SIRT6 protein levels (Figure 3B). Moreover, miR-33a also affected levels of SREBP1/2, targets of SIRT6 (Figure 3C). These findings show that SIRT6 is a target for miR-33a and suggests a negative-feedback regulation between SIRT6 and SREBP1/SREBP2. Indeed, qRT-PCR showed that


overexpression of SIRT6 in HepG2 cells resulted in a 30% decrease in miR-33a and miR-33b levels (Figure 3D). Similarly, a 40% decrease in miR-33a (the single isoform in mice) expression was detected in the liver of SIRT6-overexpressing transgenic male mice (Figure 3E). In mice, SIRT6 overexpression protects against the physiological damage induced by HFD and extends lifespan. Thus, we examined SIRT6 regulation of AMPK and SREBP in mice fed HFD and in old mice. No significant change was found in pAMPK and cleaved SREBP1 in the liver between WT and MOSES mice fed HFD (Figure S4A). However, in comparison to WT liver, the expression levels of selected SREBP1 and SREBP2 target genes, including FASN, ACC1, ACC2, LDLR, HMGCR, and HMGCS, were significantly lower in the liver of SIRT6 transgenic mice (Figure S4B). Interestingly, in 24-month-old mice, in comparison to WT, MOSES mice had increased liver pAMPK and a significant reduction in the expression levels of the selected SREBP1 and SREBP2 target genes (Figures 4A and 4B). Thus, SIRT6 maintains part of its regulation of SREBP in aged mice and in mice fed HFD. DISCUSSION

Figure 4. Effect of SIRT6 Overexpression on SREBP and AMPK Activities in Old Mice (A) Western blot analysis of pAMPK, AMPK, SREBP1, and SIRT6 in 24-monthold WT and MOSES (TG) mice. b-actin served as a loading control. Densitometric quantification of p-AMPK/AMPK is shown on the right. (n = 4 mice per group). (B) Expression profile (qRT-PCR) of fatty acid and triglyceride biosynthesis genes in the liver of 24-month-old WT and MOSES mice (n = 4 mice per group). Data represent the mean ± SEM. *p % 0.05. (C) Model of the reciprocal regulation between SIRT6 and SREBP1 and SREBP2. Upon expression of SREBP1 and SREBP2, miR-33 is cotranscribed, resulting in the downregulation of SIRT6. On the other hand, SIRT6 activates AMPK (by increasing the AMP/ATP ratio and activating LKB), which phosphorylates and inactivates SREBP1. SIRT6 also binds and deacetylates H3K56 of SREBP1 and SREBP2 promoter regions, potentially making these sites inaccessible for transcription. Moreover, SIRT6 decreases the levels of the SREBP1/SREBP2 protease complex (SCAP, S1P, and S2P), which results in diminished levels of the active forms of SREBP1 and SREBP2. See also Figure S4.

Mice overexpressing SIRT6 have lower cholesterol levels in comparison to their WT littermates. To further investigate the molecular basis for this phenomenon, we found that both HepG2 hepatoma cell cultures and liver tissues of SIRT6-overexpressing mice exhibit a significant decrease in triglyceride and cholesterol levels, indicating that these are most likely cell autonomous phenomena. In search of molecular mechanisms underlying these phenomena, we demonstrated that SIRT6 overexpression significantly reduces the levels of SREBP1 and SREBP2 and their target genes, as well as the level of the active cleaved forms of SREBP1 and 2. In parallel, SIRT6 overexpression results in increased levels of inactive phosphorylated SREBP1, AMP/ATP ratio, and LKB and AMPK activation. Finally, miR-33a and miR-33b, which are hosted within SREBP2 and SREBP1, respectively, bind to the SIRT6 transcript and reduce SIRT6 expression. Together, these findings demonstrate a reciprocal regulatory relationship between SIRT6 and SREBP1/ SREBP2 and indicate that SIRT6 is a key regulator of cholesterol homeostasis and an AMPK activator (Figure 4C). In support of this notion, we found that overexpression of SIRT6 results in at least three levels of negative regulation of SREBP1 and SREBP2: a reduction in the transcription levels of SREBP1 and SREBP2, a reduction in the active nuclear cleaved forms of SREBP1 and SREBP2 and an increase in the level of the inactive phosphorylated form of SREBP1. However, the hierarchy between the above regulatory pathways is still not clear. SREBP1 and SREBP2 are known to regulate their own transcription. Therefore, it is possible that the reduction in the transcription levels of SREBP1 and SREBP2 is the outcome of significantly reduced levels of active SREBP1/SREBP2 protein, potentially due to SIRT6-dependent AMPK activation. SIRT6/AMPK reduces SREBPs activity to a level that significantly affects their target genes (p < 0.05), whereas its effect on SREBP1c itself was almost significant (24% reduction and p = 0.061) (Figure 2G). Thus, we suggest that SIRT6’s direct


effect on SREBP mRNA levels and SREBPs autoregulation are masked by the activity of other genes. Nevertheless, the binding of SIRT6 to SREBP1/SREBP2 promoters strongly suggests that part of the reduction is due to the deacetylation of SREBP1/ SREBP2 promoter regions by SIRT6. Indeed, a significant decrease in the acetylation levels of histone H3 K56 but not K9 was found on these promoters in SIRT6-overexpressing cells versus WT (Figure 1F; data not shown). Thus, SIRT6 binding to SREBP1/SREBP2 promoters reduce their transcription levels specifically via H3 K56 deacetylation. Overexpression of SIRT6 results in significantly increased levels of liver rate-limiting enzymes in the oxidation of beta free fatty acids, HADHB, CPT1, and CROT. In addition, ACC2 levels significantly decreased. At the mitochondrial membrane, ACC2 activity results in increased malonyl-coA levels, which inhibits carnitine/palmitoyl-transferase 1 (CPT1) and reduces fatty acid b-oxidation (Abu-Elheiga et al., 2001). All together, these observations suggest an appealing model in which SIRT6 overexpression also enhances fatty acid b-oxidation. Our finding that SIRT6 activates AMPK suggests that additional SIRT6-related phenotypes may also be mediated by AMPK activation. Of the many reported AMPK-regulated pathways, we wish to highlight glucose homeostasis, cancer, and aging. Regarding glucose homeostasis, mice overexpressing SIRT6 were shown to be protected from an obesity-related decline in glucose uptake, usually associated with obesityrelated type 2 diabetes (T2D) (Kanfi et al., 2010). Similarly, in comparison to their WT littermates, SIRT6-overexpressing old mice (19–24 months) are protected against an age-related decline in glucose uptake (Kanfi et al., 2012), which is also associated with obesity-related T2D. Interestingly, metformin, an AMPK activator, is, to date, the most highly prescribed medicine used for the treatment of T2D. Thus, it would be of great interest to determine whether the protective effect of SIRT6 against T2D is partially mediated by AMPK. In a recent study, Sebastiań et al. (2012) demonstrated the role of SIRT6 as a tumor suppressor that modulates aerobic glycolysis and inhibits the Warburg effect. Similarly, Faubert et al. (2013) showed that AMPK is a negative regulator of the Warburg effect and acts as a tumor suppressor. Inactivation of AMPK in both transformed and nontransformed cells promotes a metabolic shift to aerobic glycolysis and biomass accumulation. Given this parallel together with the fact that both SIRT6 and AMPK act as tumor suppressors via hypoxia-inducible factor1a (HIF-1a), it would be of great interest to examine whether the tumor-suppressive role of SIRT6 is mediated by AMPK. To date, SIRT6 is the only sirtuin that has been shown to directly regulate mammalian lifespan. The expression levels of SIRT6 have been shown to increase in multiple tissues of rats fed with a calorie restricted diet, an energy-limited condition known to extend the rat lifespan (Kanfi et al., 2008). Moreover, male MOSES mice overexpressing SIRT6 live significantly longer than their WT littermates (Kanfi et al., 2012). Strikingly, AMPK is also activated under energy-limited conditions and, in the nematode C. elegans, AMPK activation results in long-living animals regardless of their diet (Apfeld et al., 2004). A recent study by Stenesen et al. (2013) showed that various mutations in AMP biosynthesis, which tip the balance toward an increase in the

AMP/ATP ratio, result in a significant AMPK-dependent increase in the lifespan of the fruit fly, D. melanogaster. Interestingly, we found that SIRT6 maintains part of its regulation on SREBP and AMPK in aged mice (Figures 4A and 4B). We therefore suggest that SIRT6/AMPK reduces SREBPs activity to a level that strongly affects SREBPs target genes, whereas its effect on SREBPs mRNA levels is masked by other regulators of SREBPs. Although the precise role of AMPK in mammalian longevity is still elusive, it would be of great interest to examine the extent to which the positive effect of SIRT6 on mammalian lifespan is mediated by AMPK. All together, these findings demonstrate the key role of SIRT6 as a regulator of the connection between energy homeostasis and longevity. Therefore, small molecules that can activate SIRT6 may have potential as drugs for treating age-related metabolic diseases, such as hypercholesterolemia, increased LDL cholesterol, other cholesterol-related diseases, and type 2 diabetes. EXPERIMENTAL PROCEDURES Animals Mice were kept under specific pathogen-free conditions, were raised under 12 hr day/night conditions, and had free access to standard chow diet and water. All experiments were conducted in accordance with the Institutional Animal Care and Use Committee. The Sirt6-tg/MOSES on CB6 background mice were described previously (Kanfi et al., 2010). For mouse experiments, 6-month-old male mice (WT and SIRT6 transgenic) were sacrificed, and RNA was purified from livers and processed as described in the section ‘‘Gene Expression Analysis’’ below. Aged mice were sacrificed at the age of 24 months. Mice used for HFD experiments were fed a diet with 60% of calories from fat for 16 weeks as previously described (Kanfi et al., 2010). Eight backcrosses were done with WT CB6 before the beginning of the HFD experiments. Cell Culture and Reagents The HepG2 human hepatocarcinoma cell line was grown in complete Dulbecco’s modified Eagle’s medium (low glucose) medium supplemented with 10% fetal calf serum, 1% penicillin-streptomycin mixture, and L-glutamine. For transfection, cells were grown to 50% confluency in 6-well or 10 cm plates. Transfection was performed using the lipofectamine (Invitrogen) method. The cells were harvested 48 hr after transfection. For SIRT6 transfections, 2 mg DNA (6-well plate) or 10 mg DNA (10 cm plate) was used. For siRNA transfections, cells were transfected with siRNAs (IDT) at a concentration of 50 nM, using DharmaFECT1 (Dharmacon), and harvested as indicated. MiR-33a and antimiR-33a were purchased from Applied Biosystems by Life Technologies and transfected using siPORT NeoFX transfection reagent. Lysate Preparation and Immunoblotting Cells were lysed in lysis buffer (50 mM HEPES, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, and protease inhibitor cocktail). Cleared lysates were resolved by SDS-PAGE, transferred to nitrocellulose, and incubated with primary antibodies. Immunoreactive bands were detected using ECL reagents (Pierce). For immunoblot analysis of SIRT6, monoclonal anti-SIRT6 was used (Sigma). Anti b-actin was purchased from Santa Cruz Biotechnology. Anti-SREBP1 and SREBP2 were purchased from BD Pharmingen. Anti-AMPK, p-AMPK (Thr172), ACC, p-ACC (Ser79), LKB, p-LKB (Ser428), S2P, and pSREBP1 (Ser372) were purchased from Cell Signaling Technology. S1P and SCAP antibodies were purchased from Abcam. The CaMKKb antibody was purchased from Bethyl Laboratories and the H3K56 antibody was purchased from Epigentek. Gene Expression Analysis RNA was extracted from cells with Peqgold Trifast (PEQLAB Biotechnologie) according to the manufacturer’s specifications. Afterward, from each sample,


2.5 mg was taken to a standard reverse transcription (first-strand cDNA synthesis kit, MBI Fermentas). The qRT-PCR was carried out using SYBR green fluorescent dye (Thermo Fisher Scientific), in a 15 ml total reaction volume containing 3 ml of cDNA. PCR amplification was performed with a StepOnePlus instrument (Applied Biosystems). A list of PCR primers is available on request. High-Performance Liquid Chromatography Nucleotide Analysis Sample preparation method for high-performance liquid chromatography (HPLC) analysis was adapted from the procedure of Manfredi and colleagues (Manfredi et al., 2002). Briefly, cells were suspended in 400 ml of cold 0.4 M perchloric acid in 1.5 ml Eppendorf tubes. The mixture was incubated on ice for 15 min, centrifuged at 12,000 3 g to remove the precipitate. The supernatant was neutralized (approximately 1:10) with 4 M K2CO3 solution (pH 7.5), followed by incubation for 15 min on ice. Following centrifugation (10,000 3 g), the supernatants were frozen pending chromatographic analysis. The chromatography apparatus consisted of Hitachi Elite LaChrom system equipped with diode array detector, column oven, autosampler, and a quaternary pump. All chromatographic analyses were performed at 30 C using SUPELCOSIL LC18-S HPLC Column (5 mm particle size, L 3 I.D. 25 cm 3 2.1 mm), flow rate of 0.2 ml/min under isocratic elution conditions with the following buffer composition: 50 mM potassium phosphate, 100 mM triethylamine, 0.1 mM MgCl2 (pH 6.5) (adjusted with phosphoric acid)]:Acetonitrile (98.5:1.5). Each analysis cycle was set to 30 min. The chromatographic flow was monitored at 260 nm and integrated using EZChrom Elite Software (Cichna et al., 2003; Czarnecka et al., 2005; Kehr and Chavko, 1985). Cholesterol Measurements Cells were harvested by scraping, and extracted with 200 ml of chloroform:isopropanol:NP-40 (7:11:0.1) in a microhomogenizer. The extracts were centrifuged at 15,000 3 g for 10 min. The liquid organic phase was transferred to a new tube and air-dried at 50 C to remove chloroform. Afterward the samples were put under vacuum for 30 min to remove trace organic solvent. The extracts were further analyzed using an assay kit, according to the manufacturer’s instructions (Cayman Chemical, cholesterol fluorometric assay kit). Triglyceride Measurements Cells were harvested by scraping and were homogenized in a 1 ml solution containing 5% NP-40 in water while slowly heating to 100 C. The samples were then cooled to room temperature and heated again to 100 C to ensure solubilization of all triglycerides. The samples were centrifuged for 2 min to remove any insoluble material and analyzed using an assay kit, according to the manufacturer’s instructions (Cayman Chemical, triglycerides colorimetric assay kit).

Statistical Analyses Significant differences between two groups were assessed by two-tailed Mann-Whitney test. Data are expressed as means ± SEM. Values of p < 0.05 were considered to be statistically significant. All experiments were performed at least three times.

SUPPLEMENTAL INFORMATION Supplemental Information includes four figures and one table and can be found with this article online at http://dx.doi.org/10.1016/j.celrep.2013.08.006.

ACKNOWLEDGMENTS We thank the members of the Cohen lab and Dr. Shelley Schwarzbaum for their helpful comments on the manuscript. This study was supported by the Israel Science Foundation, I-Core Foundation, Israeli Ministry of Health, and the ERC: European Research Council. H.C. is a consultant to SIRTLab, a company developing sirtuin-based therapy. Received: February 22, 2013 Revised: July 1, 2013 Accepted: August 1, 2013 Published: September 5, 2013 REFERENCES Abu-Elheiga, L., Matzuk, M.M., Abo-Hashema, K.A., and Wakil, S.J. (2001). Continuous fatty acid oxidation and reduced fat storage in mice lacking acetyl-CoA carboxylase 2. Science 291, 2613–2616. Apfeld, J., O’Connor, G., McDonagh, T., DiStefano, P.S., and Curtis, R. (2004). The AMP-activated protein kinase AAK-2 links energy levels and insulin-like signals to lifespan in C. elegans. Genes Dev. 18, 3004–3009. Brown, M.S., and Goldstein, J.L. (1997). The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 89, 331–340. Cichna, M., Daxecker, H., and Raab, M. (2003). Determination of 18 nucleobases, nucleosides and nucleotides in human peripheral blood mononuclear cells by isocratic solvent-generated ion-pair chromatography. Anal. Chim. Acta 481, 245–253. Corton, J.M., Gillespie, J.G., and Hardie, D.G. (1994). Role of the AMP-activated protein kinase in the cellular stress response. Curr. Biol. 4, 315–324.

Chromatin Immunoprecipitation Experiments were performed as previously described (Kanfi et al., 2010). Briefly, formaldehyde was added directly to cell cultures, to a final concentration of 1%. Fixation proceeded at 37 C for 10 min, and afterward cells were harvested by scraping in ice-cold PBS containing protease inhibitor. The cell pellets were resuspended in lysis buffer and then sonicated to shear DNA to segments of 500–1000 bp. After centrifugation, the chromatin solution was precleared with the addition of protein A/G agarose beads for 45 min at 4 C. Thereafter, each sample was incubated with the indicated antibody overnight at 4 C (an anti-H3 antibody was used as a positive control, and an anti-GFP antibody was used as a negative control), washed with several wash buffers to reduce background, and then eluted with NaHCO3 (50 mM). Crosslinks were reversed by addition of NaCl to a final concentration of 200 mM. Next, 20 ml proteinase K solution, 50 mM Tris (pH 7.5), and 10 mM EDTA were added to a total elution volume of 500 ml. DNA samples were purified by a PCR clean up kit (Promega). Each gene was checked up to 2,000 bp upstream of its transcription start site, in segments of 700 bp, each tested separately with a distinct set of primers in qRT-PCR. Table S1 shows the sequences of the primers that were used for the ChIP analysis.

Hardie, D.G., Salt, I.P., Hawley, S.A., and Davies, S.P. (1999). AMP-activated protein kinase: an ultrasensitive system for monitoring cellular energy charge. Biochem. J. 338, 717–722.

Gel Quantification Analysis Quantification of bands was performed using the ImageJ 1.42 densitometry software.

Hawley, S.A., Pan, D.A., Mustard, K.J., Ross, L., Bain, J., Edelman, A.M., Frenguelli, B.G., and Hardie, D.G. (2005). Calmodulin-dependent protein kinase kinase-beta is an alternative upstream kinase for AMP-activated protein kinase. Cell Metab. 2, 9–19.

Czarnecka, J., Cieslak, M., and Micha1, K. (2005). Application of solid phase extraction and high-performance liquid chromatography to qualitative and quantitative analysis of nucleotides and nucleosides in human cerebrospinal fluid. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 822, 85–90. Da´valos, A., Goedeke, L., Smibert, P., Ramı´rez, C.M., Warrier, N.P., Andreo, U., Cirera-Salinas, D., Rayner, K., Suresh, U., Pastor-Pareja, J.C., et al. (2011). miR-33a/b contribute to the regulation of fatty acid metabolism and insulin signaling. Proc. Natl. Acad. Sci. USA 108, 9232–9237. Faubert, B., Boily, G., Izreig, S., Griss, T., Samborska, B., Dong, Z., Dupuy, F., Chambers, C., Fuerth, B.J., Viollet, B., et al. (2013). AMPK is a negative regulator of the Warburg effect and suppresses tumor growth in vivo. Cell Metab. 17, 113–124. Hardie, D.G., and Hawley, S.A. (2001). AMP-activated protein kinase: the energy charge hypothesis revisited. Bioessays 23, 1112–1119.


Horton, J.D., Goldstein, J.L., and Brown, M.S. (2002). SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J. Clin. Invest. 109, 1125–1131. Kahn, B.B., Alquier, T., Carling, D., and Hardie, D.G. (2005). AMP-activated protein kinase: ancient energy gauge provides clues to modern understanding of metabolism. Cell Metab. 1, 15–25. Kanfi, Y., Shalman, R., Peshti, V., Pilosof, S.N., Gozlan, Y.M., Pearson, K.J., Lerrer, B., Moazed, D., Marine, J.C., de Cabo, R., and Cohen, H.Y. (2008). Regulation of SIRT6 protein levels by nutrient availability. FEBS Lett. 582, 543–548. Kanfi, Y., Peshti, V., Gil, R., Naiman, S., Nahum, L., Levin, E., Kronfeld-Schor, N., and Cohen, H.Y. (2010). SIRT6 protects against pathological damage caused by diet-induced obesity. Aging Cell 9, 162–173. Kanfi, Y., Naiman, S., Amir, G., Peshti, V., Zinman, G., Nahum, L., Bar-Joseph, Z., and Cohen, H.Y. (2012). The sirtuin SIRT6 regulates lifespan in male mice. Nature 483, 218–221. Kawahara, T.L., Michishita, E., Adler, A.S., Damian, M., Berber, E., Lin, M., McCord, R.A., Ongaigui, K.C., Boxer, L.D., Chang, H.Y., and Chua, K.F. (2009). SIRT6 links histone H3 lysine 9 deacetylation to NF-kappaB-dependent gene expression and organismal life span. Cell 136, 62–74. Kehr, J., and Chavko, M. (1985). Rapid determination of adenine nucleotides in brain tissue by ion-paired reversed-phase column liquid chromatography under isocratic conditions. J. Chromatogr. A 345, 267–273. Li, Y., Xu, S., Mihaylova, M.M., Zheng, B., Hou, X., Jiang, B., Park, O., Luo, Z., Lefai, E., Shyy, J.Y., et al. (2011). AMPK phosphorylates and inhibits SREBP activity to attenuate hepatic steatosis and atherosclerosis in diet-induced insulin-resistant mice. Cell Metab. 13, 376–388. Manfredi, G., Yang, L., Gajewski, C.D., and Mattiazzi, M. (2002). Measurements of ATP in mammalian cells. Methods 26, 317–326. Merrill, G.F., Kurth, E.J., Hardie, D.G., and Winder, W.W. (1997). AICA riboside increases AMP-activated protein kinase, fatty acid oxidation, and glucose uptake in rat muscle. Am. J. Physiol. 273, E1107–E1112. Michishita, E., McCord, R.A., Berber, E., Kioi, M., Padilla-Nash, H., Damian, M., Cheung, P., Kusumoto, R., Kawahara, T.L., Barrett, J.C., et al. (2008). SIRT6 is a histone H3 lysine 9 deacetylase that modulates telomeric chromatin. Nature 452, 492–496. Michishita, E., McCord, R.A., Boxer, L.D., Barber, M.F., Hong, T., Gozani, O., and Chua, K.F. (2009). Cell cycle-dependent deacetylation of telomeric histone H3 lysine K56 by human SIRT6. Cell Cycle 8, 2664–2666. Miserez, A.R., Cao, G., Probst, L.C., and Hobbs, H.H. (1997). Structure of the human gene encoding sterol regulatory element binding protein 2 (SREBF2). Genomics 40, 31–40.

Mostoslavsky, R., Chua, K.F., Lombard, D.B., Pang, W.W., Fischer, M.R., Gellon, L., Liu, P., Mostoslavsky, G., Franco, S., Murphy, M.M., et al. (2006). Genomic instability and aging-like phenotype in the absence of mammalian SIRT6. Cell 124, 315–329. Oakhill, J.S., Steel, R., Chen, Z.P., Scott, J.W., Ling, N., Tam, S., and Kemp, B.E. (2011). AMPK is a direct adenylate charge-regulated protein kinase. Science 332, 1433–1435. Raghow, R., Yellaturu, C., Deng, X., Park, E.A., and Elam, M.B. (2008). SREBPs: the crossroads of physiological and pathological lipid homeostasis. Trends Endocrinol. Metab. 19, 65–73. Sebastiań, C., Zwaans, B.M., Silberman, D.M., Gymrek, M., Goren, A., Zhong, L., Ram, O., Truelove, J., Guimaraes, A.R., Toiber, D., et al. (2012). The histone deacetylase SIRT6 is a tumor suppressor that controls cancer metabolism. Cell 151, 1185–1199. Shaw, R.J., Lamia, K.A., Vasquez, D., Koo, S.H., Bardeesy, N., Depinho, R.A., Montminy, M., and Cantley, L.C. (2005). The kinase LKB1 mediates glucose homeostasis in liver and therapeutic effects of metformin. Science 310, 1642–1646. Stenesen, D., Suh, J.M., Seo, J., Yu, K., Lee, K.S., Kim, J.S., Min, K.J., and Graff, J.M. (2013). Adenosine nucleotide biosynthesis and AMPK regulate adult life span and mediate the longevity benefit of caloric restriction in flies. Cell Metab. 17, 101–112. Viollet, B., Guigas, B., Leclerc, J., He´brard, S., Lantier, L., Mounier, R., Andreelli, F., and Foretz, M. (2009). AMP-activated protein kinase in the regulation of hepatic energy metabolism: from physiology to therapeutic perspectives. Acta Physiol. (Oxf.) 196, 81–98. Warburg, O. (1956). On the origin of cancer cells. Science 123, 309–314. Woods, A., Johnstone, S.R., Dickerson, K., Leiper, F.C., Fryer, L.G., Neumann, D., Schlattner, U., Wallimann, T., Carlson, M., and Carling, D. (2003). LKB1 is the upstream kinase in the AMP-activated protein kinase cascade. Curr. Biol. 13, 2004–2008. Woods, A., Dickerson, K., Heath, R., Hong, S.P., Momcilovic, M., Johnstone, S.R., Carlson, M., and Carling, D. (2005). Ca2+/calmodulin-dependent protein kinase kinase-beta acts upstream of AMP-activated protein kinase in mammalian cells. Cell Metab. 2, 21–33. Yang, B., Zwaans, B.M., Eckersdorff, M., and Lombard, D.B. (2009). The sirtuin SIRT6 deacetylates H3 K56Ac in vivo to promote genomic stability. Cell Cycle 8, 2662–2663. Zhong, L., D’Urso, A., Toiber, D., Sebastian, C., Henry, R.E., Vadysirisack, D.D., Guimaraes, A., Marinelli, B., Wikstrom, J.D., Nir, T., et al. (2010). The histone deacetylase Sirt6 regulates glucose homeostasis via Hif1alpha. Cell 140, 280–293.
