Circadian Measurements of Sirtuin Biology

Abstract

Many of our behavioral and physiological processes display daily oscillations that are under the control of the circadian clock. The core molecular clock network is present in both the brain and peripheral tissues and is composed of a complex series of interlocking transcriptional/translational feedback loops that oscillate with a periodicity of ∼24 h. Recent evidence has implicated NAD⁺ biosynthesis and the sirtuin family of NAD⁺-dependent protein deacetylases as part of a novel feedback loop within the core clock network, findings which underscore the importance of taking circadian timing into consideration when designing and interpreting metabolic studies, particularly in regard to sirtuin biology. Thus, this chapter introduces both in vivo and in vitro circadian methods to analyze various sirtuin-related endpoints across the light–dark cycle and discusses the transcriptional, biochemical, and physiological outputs of the clock.

1 Introduction

While the French astronomer JeanJacques d'Ortous de Mairan first noted that the diurnal leaf movements of the Mimosa plant persisted in constant darkness in the 1700s, it was not until more than 200 years later that it became established that circadian rhythms were driven by an internal endogenous circadian clock, rather than being just passive responses to a light–dark cycle. The highly conserved nature of the clock across diverse organisms, including cyanobacteria, fungi, plants, Drosophila, and mammals, suggests an evolutionary advantage to being able to synchronize behavioral and physiological processes to daily environmental cues in order to optimize energy homeostasis and enhance survival.

Many of our behavioral and physiological processes, including sleep–wake cycles, locomotor activity, hormone secretion, body temperature, and metabolism, display daily oscillations. Epidemiologic studies have revealed that many acute catastrophic events, including heart attacks, pulmonary edema, and hypertensive crises, tend to occur at specific times of day [1]. Furthermore, there is an increased risk of obesity and cardiovascular disease in shift workers, who have altered patterns of sleep/wake and fasting/feeding cycles, highlighting the close link between circadian rhythms and metabolism [2–4]. Other physiological processes, such as glucose metabolism, vary during the day. For example, humans are more glucose tolerant in the morning than in the evening, correlating with enhanced insulin sensitivity and insulin secretion in the early morning hours [5–10]. Of note, loss of rhythms in glucose metabolism is associated with the development of metabolic diseases such as type 2 diabetes [11, 12].

Although the main “driver” of circadian rhythms resides within the suprachiasmatic nucleus (SCN) of the hypothalamus in the brain, the circadian clock network is also present within most, if not all, peripheral tissues. The SCN neurons receive light input from the retinohypothalamic tract and synchronize peripheral tissue clocks through both neuronal and hormonal connections. The core molecular machinery underlying circadian rhythms within both the SCN and peripheral tissues has recently come to light through a combination of molecular and genetic studies. The core clock network consists of interwoven positive and negative transcription–translation feedback loops that function as a molecular oscillator with near 24 h precision (reviewed in ref. 13). The positive limb of the clock is composed of the basic helix-loop-helix transcription factors CLOCK and BMAL1, which heterodimerize and activate transcription of downstream targets, including the Period (Perl, 2, 3) and Cryptochrome (Cry 1, 2) genes. PER and CRY proteins then heterodimerize, translocate back to the nucleus, and bind and inhibit the activity of CLOCK/BMAL1, thereby repressing their own transcription. Additional modifying components of the core clock network have been identified, including the casein kinases I epsilon and delta, which phosphorylate PER and CRY, ultimately leading to their degradation, and the orphan nuclear receptors REV-ERBα and ROR1α, both of which are transactivated by CLOCK/BMAL1, and in turn repress and activate Bmall transcription, respectively (see Fig. 1).

Fig. 1.

Molecular overview of core clock network. The core molecular clock network consists of a series of positive and negative transcription/translation feedback loops. The positive limb consists of CLOCK/BMAL1, which heterodimerizes and transactivates expression of downstream target genes containing an E-box in their promoters, including the Period, Cryptochrome, Ror, Rev-erb, and Nampt genes. The PER and CRY proteins multimerize and inhibit the activity of CLOCK/BMAL1, until they are degraded following CKI- and FBLX3-mediated phosphorylation and ubiquitination, respectively. Rhythmic generation of Nampt results in oscillations in NAD⁺ and SIRT activity, which feeds back to regulate the positive limb of the clock

Of particular interest, several recent studies have demonstrated that NAD⁺ biosynthesis and the sirtuin family of NAD⁺-dependent protein deacetylases constitute part of a novel feedback loop within the core clock network [14–17]. CLOCK/BMAL1 directly regulate the expression of the gene encoding the rate-limiting enzyme in NAD⁺-biosynthesis, nicotinamide phosphoribosyltransferase (NAMPT), in peripheral tissues, and NAD⁺ levels oscillate across the light-dark cycle. Not only is NAD⁺ an essential cofactor for metabolic redox reactions, but it is also a substrate for SIRT1, which has been shown to deacetylate and regulate the activity of core clock proteins PER2 and BMAL1 [14, 15]. In addition, both genetic and pharmacologic modulation of NAD⁺ levels has been shown to alter clock transcriptional activity and circadian behavior and metabolism [18–20], completing a novel metabolic feedback loop within the core clock machinery. The identification of NAMPT and the sirtuin deacetylases as nutrient sensors at the nexus of circadian and metabolic pathways will likely shed additional light on how peripheral tissues coordinate fuel utilization and metabolism with the fasting/feeding and sleep/wake cycles.

Given the relatively recent finding that NAD⁺ biosynthesis and sirtuin biology are intimately connected to the circadian clock, it is important to keep the time of day in mind when designing sirtuin-related experiments. An interesting anecdote described by Ueli Schibler in his keynote lecture at the conference for the Society for Research of Biological Rhythms [21] demonstrates the significance of the time of day at which experiments are performed—a graduate student who had inherited a former postdoc's project involving DBP (albumin site D-binding protein) struggled for weeks to even detect its presence in liver, until he realized that DBP was essentially undetectable in the early morning hours (the time when he, as a “morning lark,” collected his tissue samples), while it was induced more than 100-fold by the evening (the time when the “night-owl” post-doc collected his samples). While most biological changes are not nearly as dramatic as the 100-fold difference observed in DBP levels throughout the day, even differences as small as twofold throughout the day can contribute to significant variations in experimental outcomes. Further, genomic and proteomic profiling studies have revealed that approximately 5–20% of the transcriptome or proteome displays circadian oscillations in peripheral tissues, while metabolomics studies demonstrate robust oscillations in many critical metabolites, reinforcing the importance of timing in the majority of metabolic processes [22–24]. Together, these observations underscore the importance of taking circadian timing and light cycles into consideration when designing and interpreting metabolic studies, particularly with regard to sirtuin biology. When designing such an experiment, one should ask: Has the appropriate time of day been chosen for the experiment? Have all samples been collected at the same time point? Can we gain additional physiological insight into the process by performing analyses at multiple time points throughout the day? Given the intense focus on mammalian sirtuins in the past several years, the goal of this chapter is to highlight both in vivo and in vitro methods to analyze various endpoints across the light-dark cycle, whether they are transcriptional, biochemical, physiological, or behavioral outputs of the clock. For example, given the role of the sirtuins in multiple metabolic processes, including gluconeogenesis, insulin secretion, and fatty acid oxidation, it will be important to understand the dynamics of how the sirtuins regulate these processes throughout the day at both physiological and molecular levels. Briefly, Subheading 3.1 outlines how to set up in vivo “circadian” studies in which the process of interest is analyzed over the course of 48 h in mice maintained in constant darkness, while Subheading 3.2 describes how to monitor in vivo circadian rhythms of locomotor activity as a behavioral output of the clock. Subheading 3.3 discusses how brief treatment of cultured cells with various compounds synchronizes rhythms of clock gene expression that persist for several days, thereby creating a powerful cell culture-based system to study sirtuin-related processes across the circadian cycle.

2 Materials

2.1 Circadian In Vivo Studies (48 h Constant Darkness Experiment)

1. C57BL/6J mice, aged 12 weeks (Jackson Labs).

2. Light-tight circadian cabinets with computer-controlled light–dark cycles.

3. Red light, red light-equipped room, and/or infrared nightvision goggles.

4. Standard housing mouse cages.

5. Dissection equipment (scissors, forceps, etc.).

6. Eppendorf tubes for tissue collection.

7. Liquid nitrogen.

2.2 Circadian In Vivo Studies (Locomotor Activity Rhythms)

• 1–3Same as above.
• 4Mouse cages equipped with running wheel and microswitch.
• 5Computer equipped with data acquisition and analysis software.

2.3 Circadian In Vitro Studies

1. NIH-3T3 MEF cells (ATCC# CRL-1658).

2. NIH-3T3 MEF cells infected with PER2-LUC-expressing lentivirus [25, 26] or MEFs isolated from PER2-LUC mice [27].

3. MEF culture medium: Dulbecco's modified Eagle medium (DMEM) w/4.5 g/L glucose, 10% FBS, 4 mM glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin/streptomycin.

4. 3.5 and 10 cm tissue culture dishes.

5. 37 °C incubator (C0₂-free) with LumiCycle apparatus (Actimetrics).

6. Luciferin cell culture medium: DMEM (no phenol red) plus 10 % FBS, 10 mM HEPES buffer, 0.035 % (w/v) sodium bicarbonate, 4 mM glutamine, 1 mM sodium pyruvate, penicillin/streptomycin, 0.1 mM luciferin (added immediately before use).

7. 1× sterile PBS.

8. 4 cm glass cover slips.

9. Sterile vacuum grease.

10. Forskolin.

11. Cell scrapers.

12. 1.5 ml Eppendorf tubes for cell collection.

3 Methods

3.1 Measuring Circadian Rhythms In Vivo over the Course of 48 h in Constant Darkness

For a biological process to be classified as circadian, it must display an endogenous, entrainable oscillation of approximately 24 h that is invariant across a wide temperature range. Experimentally, the rhythm must (1) repeat its cycle every 24 h for at least two cycles and (2) persist in the absence of external cues (i.e., constant darkness). Therefore, to determine whether a process is under circadian control, such as whether NAD⁺ levels are dynamically regulated by the clock, the procedure described below uses tissues that were collected for NAD⁺ measurement every 4 h over the course of 48 h in mice that have been maintained in constant darkness (12 total collection time points with 4 mice per time point = 48 mice). This approach can be applied to any process of interest, including gene or protein expression profiles, sirtuin activity, or sirtuin-related biochemistry (e.g., acetylation) or physiology (e.g., gluconeogenesis and fatty acid oxidation).

1. House C57BL/6J mice in light-tight boxes set to a 12:12 light–dark (LD) cycle in a temperature-controlled room (see ^{Notes 1} and ²). Arrange 4 mice per cage in a total of 12 cages, each of which corresponds to 1 of the 12 collection time points. Maintain animals in these conditions for a minimum of 1 week prior to the start of the experiment to allow animals to acclimate to their new surroundings.

2. Change bedding, food, and water as needed, taking care not to disturb the light conditions, as even brief exposure to altered light conditions can affect the clock (see ^{Note 3}).

3. Switch the light cycle to constant darkness (DD) 36 h prior to the first tissue collection time point, which ensures that a full circadian period has passed without entraining cues (i.e., light), but not so long that all of the individual animals become desynchronized from one another (see ^{Note 4}, Fig. 2, and Table 1).

4. 18 h prior to the first collection, remove food from the first set of mice (circadian time 0 (CT0) time point). Other mice are then serially fasted for 18 h prior to each collection time (CT4-CT44) so that each animal experiences the same length of fasting. Since fasting and food consumption are centrally regulated circadian behaviors which in turn regulate sirtuin expression and activity, fasting each mouse minimizes the effect of feeding behavior on sirtuins across the circadian cycle (see ^{Note 5}, Fig. 2, and Table 1).

5. Continue to remove food from mice from subsequent time points every 4 h, taking care not to introduce light during the food removal (see ^{Note 6}, Fig. 2, and Table 1).

6. Thirty-six hours following the switch to DD, remove the first set of 4 mice from the light-tight box and euthanize using C0₂. Euthanasia should be carried out in complete darkness (with the use of infrared goggles) or in very dim red light. After the mice have been euthanized and the light-box sealed, the lights can be turned on for the dissection (see ^{Note 7}).

7. Dissect desired tissues and collect into the appropriately labeled tubes. Tissues used for NAD⁺ measurement must be snap-frozen immediately in liquid nitrogen (details of methods to measure NAD⁺ can be found in elsewhere in this volume). Tissues for RNA and protein analysis of clock, sirtuins, or sirtuin targets can also be collected. Nuclear and mitochondrial fractions can also be collected at this time for further subcellular analysis of sirtuin activity. Continue to collect tissues every 4 h for 48 h (see ^{Notes 8}–¹⁰).

Fig. 2.

In vivo experimental timeline for 48 h tissue collection in constant darkness. Mice are released into DD at ZT12 in order to coordinate with the start of the normal dark period. Food is removed from the first set of mice at CT6, which is 18 h prior to the first tissue collection (indicated by CTO). The fasting and tissue collection is then repeated every 4 h for 44 h in order to get a full set of samples to be analyzed for rhythms of the process of interest

Table 1. Example experimental setup for 48 h tissue collection in constant darkness.

Time Point	Remove Food		Tissue Collection
CT0	Day 2	2pm	Day 3	6am
CT4		6pm		10am
CT8		10pm		2pm
CT12	Day 3	2am		6pm
CT16		6am		10pm
CT20		10am	Day 4	2am
CT24		2pm		6am
CT28		6pm		10am
CT32		10pm		2pm
CT36	Day 4	2am		6pm
CT40		6am		10pm
CT44		10am	Day5	2am

3.2 Measuring Circadian Rhythms of Locomotor Activity In Vivo

The most robust and widely used method to examine the function of the endogenous circadian clock in rodents is monitoring behavioral rhythms of locomotor activity. Here, we describe how to monitor rhythms of activity by means of wheel running, a highly precise assessment of locomotor activity given that rodents actively run on the wheels for the majority of their wake time [28]. Infrared beam crossing, an alternative to wheel running, is discussed in Subheading 4. These experiments are useful to determine whether the sirtuin-related pathway of interest within brain or periphery might affect central behavioral activity rhythms.

1. Place mice in individual cages that contain a running wheel in light-tight boxes approximately 1 h prior to the start of the active (i.e., dark) period (see ^{Note 11}). Place just a handful of bedding in the cage so as to not prevent the running wheel from moving freely. A microswitch is attached to the side of the cage and transmits a signal for every revolution of the wheel. These signals are then passed to the collection computer (see ^{Note 12}).

2. Maintain mice in 12:12 LD conditions for a minimum of 1 week to allow acclimation to the new surroundings and to collect baseline activity data (see ^{Note 13}).

3. Switch the light cycle to constant darkness and continue to collect activity data for an additional 2–3 weeks (see ^{Note 14}).

4. Activity data is recorded continuously by data acquisition and analysis software such as ClockLab (http://www.actimetrics. com). Wheel running records (i.e., actograms) are then double plotted by vertically stacking the data in overlapping 48-h traces (see Fig. 3). The following fundamental circadian parameters can then be estimated using both visual inspection and available analyses programs (see ^{Note 15}):

   1. Free-running period: The time interval between phase reference points (i.e., two peaks).

   2. Amplitude: The difference between peak and trough values of the rhythm.

   3. Phase: The timing of a reference point in the cycle (i.e., the peak) relative to a fixed event (i.e., the beginning of the night phase).

   4. Daytime activity profile: The total number of wheel running turns recorded during subjective day (or night) over the total activity during a circadian cycle or 24 h.

Fig. 3.

Example of double-plotted actogram of locomotor activity. Wheel turns are recorded in 5-min blocks and plotted as histograms along a horizontal line representing the time of day. The actogram is double plotted, meaning that 2 consecutive days are positioned next to each other both horizontally and vertically. Lighting conditions are indicated to the right. The arrow denotes the switch from 12:12 LD to constant darkness (DD) on day 10

3.3 Measuring Circadian Rhythms In Vitro

A major breakthrough in our understanding of the molecular components of the clock came with the discovery that isolated cultured fibroblasts are able to display approximately 24 h oscillations of core clock genes following brief exposure to 50 % serum, demonstrating that cells have the ability to generate cell-autonomous rhythms independent of signals from the SCN [29]. In addition to serum shock, several other compounds that activate a variety of signaling pathways have also been shown to induce circadian rhythms of gene expression in cultured cells. These include activators of protein kinase A (forskolin and butyryl cAMP), the glucocorticoid receptor (dexamethasone) [30], G-protein-coupled receptors (endothelin), as well as temperature cycles and heat shock [31, 32]. The ability to study circadian rhythms in cultured cells is a valuable tool and has numerous advantages over more timeconsuming in vivo studies, including the ease in which cells can be genetically manipulated to either overexpress or knockdown clock-or sirtuin-related genes. Here, we outline an experiment in which MEF cells are pharmacologically synchronized using forskolin to generate coordinated ∼24 h oscillations of the molecular clock transcription factors and their downstream metabolic targets. Rather than synchronization of all of the cells at one time point, followed by a continuous collection of samples every 4 h for 44 h, we describe here a more convenient method whereby cells are synchronized at four time points over 36 h and then collected at three time points on the final day of the experiment (see Fig. 4 (Method 3) and Table 2). In addition, we describe a method to simultaneously monitor circadian rhythmicity during the time course experiment using a biological circadian clock reporter construct, Period2-luciferase (Fig. 5) [27]. Various outputs, including rhythms of gene or protein expression, NAD⁺, or other biochemical or physiological endpoints of interest, can then be examined in the samples collected from the WT-MEFs.

Fig. 4.

In vitro experimental timelines for synchronization and 48 h collection time points in cells. Three different experimental setups for synchronizing cells in culture are presented. Method 1 entails synchronization of all cells at one time point, followed by collection of samples every 4 h over the course of 48 h. Method 2 entails synchronization of the cells every 4 h, enabling a single collection time point (which may be necessary if a particular assay requires all samples be run at the same time). Finally, Method 3 (described in text) outlines a convenient approach whereby cells are synchronized at four time points over 36 h and then collected at three time points on the final day of the experiment such that no overnight stays are required

Table 2. Example experimental setup for synchronization and 48 h collection in cells.

Time Point	Synchronization Time		Collection Time
44	Day 2	8pm	Day 5	4pm
40				12pm
36				8am
32	Day 3	8am		4pm
28				12pm
24				8am
20		8pm		4pm
16				12pm
12				8am
8	Day 4	8am		4pm
4				12pm
0				8am

Fig. 5.

Monitoring PER2-LUC oscillations in vitro. MEFs can be isolated from either PER2-LUC mice or can be infected with PER2-LUC lentivirus. Cells are then synchronized by the addition of forskolin directly to the media, and bioluminescence can be monitored for the remainder of the experiment in a Lumicycler machine. Each plate should display ∼24 h rhythms of luminescence beginning from 12 to 24 h after forskolin treatment. An example trace is shown

1. Passage MEF cells in 10 cm dishes according to ATCC instructions (see ^{Note 16}). MEF cells that have been stably infected with PER2-LUC-expressing lentivirus can also be plated in parallel, allowing for real-time bioluminescence monitoring in order to assess synchronization (see ^{Note 17}).

2. One day prior to synchronization, passage 12 near-confluent (∼90 %) 10 cm dishes of MEFs into twenty-four 10 cm dishes in normal MEF culture media (12 time points × 2 duplicates =24 plates) and plate four 3.5 cm dishes of the PER2-LUC MEFs, such that all dishes are at least 70 % confluent at the start of the experiment the following day (∼1.2 × 10⁶ cells per 10 cm dish, ∼2 × 10⁵ cells per 3.5 cm dish).

3. For the 10 cm dishes, add forskolin directly to media in each dish at the designated time (see Fig. 4 and Table 2) to a final concentration of 10 μM. Incubate at 37 °C for 30 min, wash once with 1× PBS, replace with fresh MEF culture medium, and return to 37 °C until the time of collection (continue to step 5) (see ^{Notes 18} and ¹⁹).

4. For the 3.5 cm dishes, add forskolin directly to media in each dish at the designated time (see Fig. 4 and Table 2) to a final concentration of 10 μM and incubate at 37 °C for 30 min. Wash once with 1× PBS and add 1.2 ml luciferin-containing medium. Coat the top rim of the plate with a thin layer of sterile vacuum grease and place cover glass on top to prevent evaporation. Gently press on cover glass to create a seal and place plates in the Lumicycler (maintained at 37 °C without CO₂) in order to monitor bioluminescence for the remainder of the experiment (see Fig. 5). Each plate should display ∼24 h rhythms of luminescence beginning from 12 to 24 h after forskolin treatment (data analysis software (Clocklab) is available with the Actimetrics Lumicycle apparatus). In addition, plates synchronized at different times should display similar period lengths, amplitudes, and be in phase with one another (i.e., peaks and troughs should occur at similar times post-synchronization).

5. For the 10 cm dishes, aspirate media and wash cells once with PBS at the designated collection time points (see ^{Note 20}). Add 1.5 ml PBS, scrape cells and collect in microcentrifuge tubes, and centrifuge at 1,000 × g for 2 min. Aspirate supernatant and resuspend cells in 1 ml PBS, repeat spin, and freeze cell pellets at −80 °C. Depending on the goals of the experiment, various endpoints can then be examined, including RNA, protein, biochemical analyses, or NAD⁺ quantification. For NAD⁺ analysis, an aliquot of the cells can be set aside prior to freezing for protein quantification in order to normalize for any difference in cell number between samples. NAD⁺ measurement is described elsewhere in this volume.

6. If PER2-LUC bioluminescence monitoring is not performed, then it will be necessary to confirm proper synchronization by the analysis of transcript levels of canonical clock genes. RNA should be extracted from the frozen cell pellets, reverse transcribed, and the resultant cDNA should be amplified by RT-PCR with specific primers for canonical clock genes, including Clock, Bmall, Per1-3, Cry1-2, Rev-Erbα, Rorα, and Dbp. Rhythms of expression of other genes of interest, including Nampt, the sirtuins, and downstream targets can also be analyzed in parallel.