Overexpression of Sir2 in the adult fat body is sufficient to extend lifespan of male and female Drosophila

Sir2 is the most intensively discussed longevity gene in current aging research. Although, the gene encoding for a NAD+-dependent histone deacetylase initially was found to extend lifespan of various organisms ranging from yeast to mammals, serious doubts regarding its role in longevity have been expressed recently. In this study, we tested whether tissue-specific overexpression of Sir2 in the adult fat body can extend lifespan when compared to genetically identical controls. We also wanted to elucidate the mechanisms by which fat body Sir2 promotes longevity by studying the phenotypic and transcriptional changes in the fat body. We found that moderate (3-fold) Sir2 overexpression in the fat body during adulthood only can promote longevity in both sexes by roughly 13 %. In addition, we obtained transcriptional profiles elicited by this overexpression and propose a role for Sir2 in lipid droplet biology especially under conditions of starvation. Furthermore, our data do not support the idea of Sir2 mediating the response to dietary restriction (DR) because transcriptional profiles of fat bodies after DR or Sir2 overexpression do not match. This study provides additional independent evidence for the concept of Sir2 as a longevity gene and as a promising pharmacological target to cure age-related diseases.


INTRODUCTION
In current aging research, Sir2 is one of the most prominent longevity genes and at the same time the most intensively discussed one. Kaeberlein et al. described Sir2 (silent information regulator 2) as a NAD + -dependent histone deacetylase that extends replicative lifespan of yeast [1]. It is highly conserved in eukaryotes and lent its name to the family of Sirtuins (Sir2-like genes), which has five members in Drosophila (Sir2, Sirt2, Sirt4, Sirt6 and Sirt7) and seven members in mammals.
Findings in C. elegans and Drosophila confirmed a role of Sir2 on lifespan. Tissenbaum and colleagues showed that an increased copy number of Sir2.1 extends lifespan of worms by up to 50 % [2]. In Drosophila, ubiquitous, constitutive overexpression as well as pan-

Research Paper
neuronal overexpression only in the adult animal was shown to extend lifespan [3,4]. Studies in mice have demonstrated that the Sir2 homolog Sirt1 does not increase life expectancy but promotes a longer health span by prevention of age associated diseases [5][6][7].
Due to the apparent evolutionarily conserved effect of Sir2/Sirt1 on health and ageing, the physiological functions and protein targets of the enzyme have been intensively studied in mammals in order to cure agerelated diseases. SIRT1 was shown to increase insulin sensitivity and regulate energy expenditure through stimulation of oxidative metabolism like fatty acid oxidation and oxidative phosphorylation in mice [5,[8][9][10]. This predestines the enzyme as a novel target for the treatment of diabetes [11]. In addition, it has been implicated in the etiology and/or cure of cancer, Alzheimer's and other age-related diseases [12][13][14]. In contrast to mammals, in Drosophila only few studies tried to elucidate physiological functions of Sir2 and the molecular mechanisms, which may mediate its effect on life span. A screen for obesity-inducing genes in Drosophila larvae pointed to a role for Sir2 in regulating fat metabolism and response to amino-acid starvation [15]. Moreover, Sir2 apparently regulates expression of genes involved in fat metabolism and the lack of Sir2 increases fat deposition under normal conditions and consequently impairs starvation survival of flies [16].
It has been hypothesized that Sir2 mediates the effects of dietary restriction (DR) on fly lifespan because decreased activity of the histone deacetylase RPD3, which is downregulated in DR, increases SIR2 activity [3]. This hypothesis is corroborated by the recent results that revealed a diet-dependent effect of Sir2 on lifespan of Drosophila [17]. In addition, P53 is discussed as a downstream target of RPD3 and SIR2 in mediating longevity [4].
Strong doubts over the lifespan prolonging effect of Sir2 overexpression have recently been expressed [18]. As different genetic backgrounds of control and experimental animals can be a confounder in lifespan experiments, Burnett et al. [19] aimed to validate results for C. elegans and Drosophila with genetically matched controls and could not confirm the positive effect on lifespan. However, a very recent study has supported the initial observation by showing that genetic manipulation of Sir2 in the fly's fat body elongates lifespan in Drosophila [17]. Thus, an urgent need to either confirm or rebut these effects of Sir2 expression levels on lifespan exists. We used the inducible GeneSwitch Gal4/UAS system [20] to drive Sir2 overexpression in the fat body only, which offers two major advantages: first, it is possible to restrict overexpression to the adult animals without affecting development and second, Sir2 overexpressing flies and control flies have the same genetic background since overexpression is induced by treatment with mifepristone (RU486). We found that Sir2 overexpression in the fat body only can promote longevity in both sexes. Furthermore, a detailed analysis of the fat tissue revealed no overt structural changes. However transcriptional profiles of fat bodies after DR or Sir2 overexpression suggest that the interventions promote longevity through independent mechanisms.

Moderate Overexpression of Sir2 in the Fat Body Elongates Lifespan of Drosophila
To avoid differences in genotypes between control and experimental animals, we used the inducible GeneSwitch System [20] to overexpress Sir2 in the fat body of adult flies. Here, tissue-specific overexpression of Sir2 is achieved by administration of mifepristone (RU486) to the F1 generation of a cross of UAS-Sir2 with the driver line S 1 106-Gal4 while genetically and developmentally identical controls feed on normal medium supplemented with the same dose of the diluent ethanol. We first examined the expression level of Sir2 under the experimental conditions. UAS-Sir2>S 1 106-Gal4 flies were fed 400 µM RU486 or diluent only for 24 h. Fat bodies were dissected and qPCR for Sir2 mRNA was performed. We detected an approx. 3.5 fold and 2.5 fold up-regulation of Sir2 gene expression in the fat bodies of females and males, respectively (Fig. S1). This system induces a moderate increase of Sir2 expression, which is presumably in the physiological range of expression changes.
Until recently, studies that aimed to investigate the effect of Sir2 on lifespan mainly manipulated its gene expression in whole animals or the nervous system. We decided to study the effect of overexpressing Sir2 in the fat body as it is strongly expressed in this tissue [3]. Moreover, the fat body integrates the response to both immune and metabolic stimuli, which are both important modulators of the aging process [21,22]. To do so, we determined lifespan of mated UAS-Sir2>S 1 106-Gal4 flies fed on medium with or without RU486 throughout adult life. Survival curves of females ( Fig. 1A) and males (Fig. 1C) overexpressing Sir2 are significantly shifted to the right. As shown in Table 1, median lifespan of females and males treated with RU486 is increased by 12 % and 13 % compared to controls, respectively. These data show that even the relatively weak overexpression of Sir2 induced by the S 1 106-Gal4 line in the adult fat body promotes longevity in flies, largely by delaying the onset of cohort aging. To rule out the possibility that the GAL4 activator RU486 itself may elongate lifespan of Drosophila, we performed lifespan experiments with flies from a crossing of the w 1118 strain and the driver line S 1 106-Gal4. Flies were treated as described above and no significant shift in survival curves was observed for either females (Fig. 1B) or males (Fig. 1D) compared to controls.

Gene Expression Profiles Induced by Sir2 Overexpression and Dietary Restriction Do not Match
The finding that fat body Sir2 can promote a long life led us to examine which candidate mechanisms may play a role in conferring Sir2-dependent longevity. As several studies suggest that Sir2 mediates at least parts of the response to dietary restriction (DR) [3,17] we www.impactaging.com aimed to elucidate how these two interventions alter gene expression patterns of the fat body and to what extent transcriptional responses match. We performed microarray analysis with RNA isolated from fat bodies of mated, age-matched females either subjected to DR for 7 d or with fat body specific Sir2 overexpression after 24 h induction by RU486. The control and DR food was prepared according to recommendations by Ja et al. [23] to prevent desiccation, which robustly result-ed in prolonged lifespan of male and female w 1118 flies (Fig. S2). Surprisingly, when comparing gene expression profiles we noticed that only roughly 10 % of up-(p value 0.355) or down-regulated genes (p value 0.383) overlap between both samples (Fig 2A, B). This is a surprising small overlap, indicative for almost completely independent processes, especially when taking into account that fat tissue of age-and gendermatched flies was used for both analyses.   Table  1, n = 100 for each group. www.impactaging.com Data derived from log rank analysis of the survival curves shown in Fig. 1. The maximum lifespan is the median of 10 % survival. Log rank test was performed using GraphPad Prism 6.00. ns, not significant. Experiments shown are representatives of two independent experiments.  [25]. The Venn diagram shows considerable intersections between these different sample types (statistical analyses were performed with Fisher's exact test; see Table 2 for Furthermore, hierarchical clustering of the microarray data shows substantial differences in gene expression patterns induced by both interventions (Fig 2C).
Especially genes down-regulated in DR sets are upregulated in Sir2 gene sets. Additionally, we did not detect an increase of Sir2 transcription in response to DR as reported by other groups, neither in whole animals nor in their fat tissues (Fig. 2D). Taken together, these results imply that fat body Sir2 evokes other cellular processes than those elicited by DR.
Gene IDs of proteins which are associated with lipid droplets [25] and of genes which were differentially regulated in 2 of 3 microarrays of fat bodies of female or male UAS-Sir2>S 1 106-Gal4 flies after feeding of RU486. www.impactaging.com

Fat Body Sir2 Regulates Lipid Droplet Proteome Associated Genes in a Gender-specific Manner
Based on the transcriptome studies performed with isolated fat tissue from Sir2 overexpressing males and females, we focused on those genes that are relevant for fat tissues. Within this tissue lipid droplets are of central importance to regulate fat utilization [24]. Thus, we hypothesized that fat body Sir2 modulates the composition of the lipid droplet proteome and compared the gene expression data from fat bodies overexpressing Sir2 with a list of Gene IDs corresponding to the lipid droplet proteome as defined by Beller et al. [25].
Corresponding genes of 39 out of 254 lipid droplet associated proteins were found to be differentially regulated in fat bodies overexpressing Sir2 which corresponds to 15 % of the lipid droplet proteome ( Fig.  3A; p value 0.007). The (putative) functions of these genes are very diverse and range from lipid storage and fatty acid metabolism, stress response and energy metabolism, to chromatin assembly ( Table 2).
When comparing gene expression patterns of male and female fat bodies overexpressing Sir2 we found great differences between genders. Only 40 of 619 genes upregulated in females were also found within the 320 genes up-regulated in males (Fig. 3B, p value 0.123). This overlap corresponds to 6.5 % and 12.5 % of upregulated genes in females and males, respectively. The situation was different for down-regulated genes where 11 % and 22 % were overlapping between females and males, respectively (p value < 0.0001). These gender-specific differences are mirrored in the comparison of genes affected by Sir2 overexpression and the genes corresponding to proteins associated with lipid droplets. Considering the data for males and females separately 17 (7 %) and 25 (10 %) of the lipid droplet proteome genes are regulated, respectively (Fig. 3A). But only 1 gene (Glutamate oxaloacetate transaminase 2) is up-regulated in both (Table 2) whereas all other genes are not or inversely regulated in the respective other gender. Taken together, these results point to a role for fat body Sir2 in modulating the composition of the lipid droplet proteome, which is obviously gender-specific.

Fat Body Sir2 Increases Starvation Resistance in Male Flies
The fat body is the major storage organ of the fly, which regulates the response to differing nutrient conditions. Sir2 is capable of sensing cellular energy levels, as its enzymatic activity is NAD + -dependent [26]. We hypothesized that fat body Sir2 participates in regulating the response to changes in nutrient conditions when promoting longevity. One measure for the ability to react to changing nutrient supply is to determine starvation resistance. Mated UAS-Sir2>S 1 106-Gal4 flies fed on 1 % agar with or without RU486 and starvation medium was changed every 2-3 days. Dead flies were counted 1-3 times a day. Fat body Sir2 appeared to have no impact on female starvation resistance as median lifespan of control and Sir2 overexpressing flies showed no difference (Fig. 4A, C). However, fat body Sir2 increased starvation survival in males by 13 % (Fig. 4B, C). In line with the transcripttome analysis we found a small gender-specific reaction to starvation after Sir2 overexpression in the fat body.

No Influence of Sir2 on Fat Storage and Fat Cell Morphology
As Sir2 overexpression in the fat body did not only extend lifespan of the flies but also influenced the relative abundance of lipid droplet proteome associated genes and starvation resistance in a gender-specific manner we wanted to elucidate the role of fat body Sir2 on fat storage and morphology of the fat cells. We tested if Sir2 overexpression affects relative body fat content, body weight, the morphology of the fat cells and the size of the fat droplets in males and females.
Mated UAS-Sir2>S 1 106-Gal4 males or females were fed on medium containing RU486 or ethanol for 7 d before they were weighed and relative body fat content was quantified by a coupled colorimetric assay according to Hildebrandt et al. [27]. Neither sex showed fat body Sir2 -dependent effects on fat storage (Fig.  5A) and body weight (Fig. 5B). To analyze changes in the morphology of the fat body we performed Bodipy staining with L3 larvae (Fig. 5 D, G), adult females ( Fig. 5 E, H) and males (Fig. 5 F, I). Although fat cell size and filling levels of fat stores obviously varied substantially between all sample types, no effect of Sir2 overexpression (  www.impactaging.com

DISCUSSION
The aim of this study was to test whether tissue-specific overexpression of Sir2 in the fat body of adult flies can elongate lifespan and to elucidate the mechanisms by which fat body Sir2 promotes longevity. We showed that moderate (roughly 3-fold) Sir2 overexpression in the fat body can extend median lifespan of both male and female Drosophila melanogaster by about 13 %. With this study, we provide additional independent evidence for the concept of Sir2 as a longevity gene in Drosophila as suggested by earlier studies [3,4,16,17]. The UAS-line used in our study has not been employed for lifespan experiments so far and does not show increased expression of dnaJ-H, which has been shown to affect apoptosis control of Sir2 and may also interfere with Sir2 functions with regards to longevity [28]. Importantly, we used the inducible GeneSwitch Gal4 System; this removes the use of different genetic backgrounds in control and Sir2 overexpressing flies, a potentially severe confounder in lifespan studies. Geneswitch also allows overexpression during adulthood only, thus removing any developmental testing potential gerontogenes from a therapeutic perspective. However, our results contradict the study from Burnett et al. [19] who did not observe an effect of ubiquitous overexpression of Sir2 on Drosophila lifespan. This may be due to antagonistic effects on lifespan caused by Sir2 overexpression during development, and/or due to the site of overexpression; by restricting overexpression of Sir2 to the fat body, possibly harmful effects caused by overexpression in other tissues may be excluded. In this context it is notable that overexpression of Sir2 in the nervous system prolonged lifespan of flies [3]. These results are in line with the observation that only these two tissues express SIR2 protein at high levels [3].
We also aimed to test different hypotheses of how fat body Sir2 may confer longevity. Our first working hypothesis was to presume that Sir2 overexpression mimics dietary restriction, which has been proposed by other studies [3,17,29]. This led us to compare microarray data generated from fat bodies overexpressing Sir2 with those obtained from fat bodies of animals subjected to DR. As the induced transcriptional changes of both interventions were different our data did not confirm this hypothesis. Instead they support the alternative hypothesis that DR and Sir2 overexpression are indendent processes. Furthermore, dietary restriction had no effect on Sir2 mRNA levels in fat tissue. Our results join a canon of studies presenting conflicting results regarding this topic. Previous findings in yeast [30][31][32][33], flies [3,17,19] and rodents [34,35] are inconsistent, which may be due to application of different dietary /caloric restriction regimes, different sites of ectopic expression or knockout of Sir2 and/or genetic backgrounds. We used the food preparations recommended by Ja et al. [23] to circumvent adverse side effects caused by desiccation of the flies, which robustly produced a DR response in both sexes and the food used in the Sir2 experiments was identical except for addition of RU486 dissolved in ethanol or ethanol alone. However, the data presented here suggest that fat body Sir2 does not mediate the effects of DR on lifespan of Drosophila.
It has also been proposed that Sir2 in the fat body plays a role in regulating fat storage and mobilization as Sir2 / Sirt1 has been implicated in regulation of fat metabolism in flies and mammals. Firstly, we compared the part of the fat body's transcriptome that is altered by Sir2 overexpression with those genes coding for the lipid droplet proteome as these organelles are known to be central players for the control of fat storage and mobilization [36][37][38][39]. A considerable overlap between these data sets suggested a role for fat body Sir2 in modulating the composition of the lipid droplet proteome, a hypothesis that is supported by the statistical data. Nevertheless, as lipid droplets have been shown to regulate aging processes [40] our data provide a good starting point to unravel the putative role of Sir2 in lipid droplet biology. The overlap with genes coding for lipid droplet proteins was highly sex specific, a feature that was also seen for starvation resistance and the Sir2 induced transcriptional changes in fat tissues. Our findings that fat storage, body weight and lipid droplet size were not changed after Sir2 overexpression, contrast with other recent studies which demonstrated an impact of RNAi-mediated knockout of Sir2 in the fat body or constitutive whole body knockout, respectively, on fat metabolism [16]. Furthermore, data from mouse studies indicate a role for Sir2 in the regulation of lipid metabolism [8,10]. These divergences may result from the different expression system used in our study. We aimed to overexpress Sir2 only moderately in the fat body and nowhere else, and only during adulthood. This strategy should have only minimal confounding effects on e.g. fat storage caused by complete silencing of Sir2 in either the fat body/liver or the whole animal.
Overall, we conclusively showed that Sir2 is a longevity gene in Drosophila. With this study we provide additional independent evidence for its ability to extend lifespan of male and female flies when moderately overexpressed in the adult fat body. We also obtained transcriptional profiles elicited by this overexpression and propose a role for Sir2 in lipid droplet biology especially under conditions of starvation. However, the precise conditions under which Sir2 affects fat www.impactaging.com utilization and the molecular mechanisms involved in this function remain to be elucidated. Furthermore, our transcriptional data do not support the idea of Sir2 mediating the response to dietary restriction. This study again highlights the relevance of Sir2 for aging processes and its potential as a pharmacological target to cure age-related diseases. Microarray Analysis. Microarray analyses were essentially performed as described earlier [41,42]. In brief, cDNA synthesis from RNA isolated from fat bodies was performed with Prime Script RT (Takara) using the following primers: OdT T7 I (5'-GAG AGA   GGA TCC AAG TAC TAA TAC GAC TCA CTA  TAG GGA GAT TTT TTT TTT TTT TTT TTT TTT T  G/A/C-3') and CapFinder Sp6rG (5'-CAG CGG CCG  CAG ATT TAG GTG ACA CTA TAG A rGrGrG-3').  cDNA was amplified with OdT T7 II (5'-GAG AGA  GGA TCC AAG TAC TAA TAC GAC TCA CTA  TAG G-3') and Adaptor Sp6rG (5'-GAC GCC TGC  AGG CGA TGA ATT TAG G-3')  BODIPY Staining and Determination of Diameters of Lipid Droplets. After feeding on RU486 or control medium for 24 h L3 larvae or adults generated by crossing UAS-Sir2 flies to the driver line S1106-Gal4 were dissected, fixed and fat bodies were stained with a solution containing 1 µg/ml BODIPY (Invitrogen) and 5 µg/ml Hoechst (Sigma) for 30 min at room temperature. Slides were mounted in Mowiol/DABCO (Roth) and photographed with a Zeiss Axio imager 7.1 microscope equipped with a Zeiss Axiocam MRM camera within the next two days. 14 photographs taken from 7 biological replicates were analyzed per group with AxioVision 4.1 (Zeiss). To measure fat droplet diameters a line was drawn across each picture and only droplets intersecting with the line were measured using a measurement tool implemented in the program.

METHODS
Starvation Resistance. Mated Sir2>S 1 106-Gal4 flies were kept on 1 % agar containing either 400 µM RU486 or ethanol and vials were changed every 2-3 days. Dead flies were scored 1-3 times a day. Statistics and Bioinformatics. All statistics were performed using GraphPad Prism version 6.00 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com. Lifespan and starvation survival data were analyzed by log rank test and data from qPCR, body fat quantification, body weight and diameters of lipid droplets by Mann-Whitney-U Test. Statistical analyses of Venn-diagram data were performed with Fisher's exact test.
has a negative effect on organismal growth and induces a peculiar set of target genes. BMC Genomics.    Table S2. Data derived from log rank analysis of the survival curves shown in Fig. S2. The maximum life span is the median of 10 % survival. Log rank test was performed using GraphPad Prism 6.00. Experiments shown are representatives of two independent experiments. ns, not significant. www.impactaging.com