Relationship between circadian oscillations of Rev-erbα expression and intracellular levels of its ligand, heme

Abstract

The nuclear hormone receptors, REV-ERBα [NR1D1] and REV-ERBβ [NR1D1], were recently demonstrated to be receptors for the porphyrin, heme. Heme regulates the ability of these receptors to repress transcription of their target genes via modulation of the affinity of the receptor's ligand binding domain for the corepressor, NCoR. The REV-ERBs function as critical components of the mammalian clock and their expression oscillates in a circadian manner. Here, we show that in NIH3T3 cells intracellular heme levels also oscillate in a circadian fashion. These data are the first to show the temporal relationship of intracellular heme levels to the expression of its receptor, Rev-erbα, and suggest that the rapid oscillations in heme levels may an important component regulating REV-ERB transcriptional activity.

Nuclear hormone receptors (NHRs) regulate many physiological processes ranging from growth and differentiation to metabolism and reproduction [1]. NHRs function as ligand regulated transcription factors and function as receptors for a variety of hydrophobic ligands including steroid hormones, thyroid hormones, and dietary lipids. Many of the 48 human NHRs are still characterized as orphan receptors due to the lack of an identified ligand [2]. Two of these receptors, which until recently were characterized as orphan receptors (REV-ERBα [NR1D1] and REV-ERBβ [NR1D2]), are particularly interesting NHRs due to their role in regulation of both metabolism and the circadian rhythm [3,4].

Circadian rhythms are common in physiology and are essential for normal regulation of an array of process including metabolism, thermoregulation, blood pressure, renal function and the sleep-wake cycle. At the cellular level, the circadian rhythms are generated by feedback loops in the expression of clock genes where heterodimers of BMAL1 and CLOCK induce the expression of Crytochrome (Cry) and Peroid (Per) genes [5]. When CRY and PER reach a crucial threshold they repress the stimulatory effects of the BMAL1/CLOCK heterodimers on expression of their own genes. REV-ERBα (a transcriptional repressor) and another NHR, RORα [NR1F1] (a transcriptional activator) play critical roles in regulating the cyclic expression of Bmal1 and are thus important components of the mammalian clock [6-8]. Additional feedback is contributed by the BMAL1/CLOCK heterodimer directly regulating Rev-erbα expression via an E-box within its promoter [9,10]. Rev-erbα^−/− mice express Bmal1 in an abnormal fashion and the mice also exhibit alterations in the phase and period of their circadian behavior patterns [7].

Recently, the porphyrin heme was identified as a ligand for REV-ERBs [11,12]. Heme binds directly to the ligand binding domain of either REV-ERB and regulates the ability of the receptors to repress its target genes including Bmal1[11,12]. Interestingly, the level of expression of the rate limiting enzyme in heme production, δ-aminolevulinic acid synthase 1 (Alas1), is also regulated in a circadian fashion [13,14]. Components of the mammalian clock including NPAS2/BMAL1 and PER have been shown to be essential regulators of the circadian expression pattern of Alas1 [13,14]. Although the circadian expression pattern of Alas1 has been characterized, there have been no studies examining intracellular heme levels in circadian synchronized cells.

Materials and Methods

Cell culture

NIH3T3 cells were maintained in DMEM + 10% FBS as per ATCC guidelines. Twenty-four hours prior to synchronization, cells were plated in 6-well plates at a density of 10⁶ cells/well. Cells were synchronized by a 50% horse serum shock [15-17] or 30 μM hemin shock [14] for 2 h. Following the 2 h shock, cells were harvested for RNA (n = 4) every 4 h for 52 h. At the timepoints described for RNA isolation, cells were harvested for heme measurements (n = 4) as described below.

Quantitative RT-PCR

Quantitative RT-PCR was performed as previously described [18,19]. Primers have been previously described [20].

Heme measurements

Intracellular heme was measured as previously described [11].

Statistics

One-way ANOVA was utilized to analyze the alterations in gene expression and heme concentrations as a function of time. The Bonferroni test was utilized to assess significant (P < 0.05) differences between various times. A point was deemed significant if it was significantly different than the nadir point of the trough either just before or after the peak.

Results

Serum shock has been demonstrated synchronize the expression of mammalian clock genes in cultured cells revealing the diurnal pattern of expression of Per2, Bmal1, and Rev-erbα among other genes [15-17]. We shocked NIH3T3 cells with 50% horse serum for 1 h and then examined the expression of Per2, Bmal1, and Rev-erbα by qPCR every 4 h for 2 days post serum shock. As illustrated in Fig. 1 Per2 and Bmal1 expression displays a circadian pattern with opposing phases as expected indicating that the serum shock successfully synchronized the cells. Per2 expression varied as much as 40-fold (Fig. 1A) while Bmal1 expression varied approximately threefold (Fig. 1B). Both displayed the expected period of approximately 24 h. Rev-erbα expression also displayed the expected circadian pattern of expression with an approximate 22-fold variation in expression (Fig. 2A). Interestingly, when we monitored the intracellular levels of heme in NIH3T3 cells we observed rapid oscillations in heme concentrations with a frequency much greater (period∼8 h) than the 24 h period normally observed for clock genes (Fig. 2B). Heme concentrations varied within the range of 0.8 to 2.3 μM/mg protein, which is well within the range found to modulate REV-ERB target genes such as Bmal1 as much as threefold in HepG2 cells (0.7 to 1.5 μM/mg protein) [11]. This pattern was maintained for ∼32 h before the oscillations appeared to dampen, possibly due to desynchronization within the cultured cells. No oscillations in heme levels were observed in cells that had not been synchronized with the 50% horse serum shock (data not shown).

Fig. 1.

Circadian oscillations of Per2 and Bmal1 expression in NIH3T3 cells following serum shock. Cells were synchronized with a 50% horse serum shock followed by analysis of Per2 and Bmal1 mRNA by qPCR at the times indicated. *P < 0.05.

Fig. 2.

Circadian oscillations of Rev-erbα expression and intracellular [Heme] in NIH3T3 cells following serum shock. Cells were synchronized with a 50% horse serum shock followed by analysis of Rev-erbα mRNA by qPCR and intracellular [Heme] at the times indicated. *P < 0.05.

A heme shock (30 μM hemin) has also been shown to have the ability to synchronize the circadian rhythm in cultured cells [14]. We also performed a heme shock and measured Per2 expression to monitor synchronization of the cells as had been performed previously [14]. We noted circadian oscillations in Per2 expression that were very similar to those noted by Kassik and Lee [14] (Fig. 3A). We also monitored intracellular heme levels in NIH3T3 cells that had been synchronized with the 30 μM heme shock. As illustrated in Fig. 3B, intracellular heme levels were initially quite elevated compared to the concentration range noted in the serum shocked cells (∼5-fold greater). This is consistent with the elevated levels of heme oxygenase-1 expression that we noted in these cells (data not shown). As the intracellular heme levels decreased to levels similar to non-heme shocked cells we noted what appeared to be oscillations similar to those observed in the serum shock; however, multiple significant oscillations were not achieved. (Fig. 2B).

Fig. 3.

Circadian oscillations of Per2 expression and intracellular [Heme] in NIH3T3 cells following heme shock. Cells were synchronized with a 30 mM hemin shock followed by analysis of Per2 mRNA by qPCR and intracellular [Heme] at the times indicated. *P < 0.05.

Discussion

Although circadian oscillations in the expression of the rate limiting enzyme for heme production, Alas1, have been previously noted [13,14] no direct observation of oscillations of intracellular heme levels have been observed until now. The fact that Alas1 oscillations have been observed in multiple models along with our observation of intracellular heme oscillations suggests that this may be a common physiological phenomenon. The role heme plays in the circadian rhythm is unclear, but heme coordinating protein components of the circadian oscillator such as NPAS2 have been described [21,22]. After we and others found that REV-ERBα and REV-ERBβ act as heme receptors [11,12], we sought to determine the relative levels of intracellular heme during the cellular circadian rhythm following synchronization with a serum shock. Both RORα and Rev-erbα exhibit circadian patterns of expression with a period of approximately 24 h; however, they are 180° out of phase with one another. They often compete for the identical DNA response element as exemplified by the elements within the Bmal1 promoter and since RORα is a constitutive activator of transcription and REV-ERBα is a constitutive repressor of transcription the oscillations in expression of these two transcription factors may underlie the oscillations in observed Bmal1 expression [3,23]. However, we now know that the repressor activity of REV-ERBα requires heme binding and based as demonstrated in this study, intracellular heme levels also oscillate. Since intracellular heme levels oscillate at a much higher frequency than that of REV-ERBα, this indicates that it is likely that the transcriptional repressor activity of REV-ERBα varies even during the time when it is the dominant factor occupying the REV-ERBα/RORα response elements such as those found in the Bmal1 promoter (Fig. 4). The frequency of heme oscillations may vary, and the high frequency oscillations that we observed may be specific to the cell-type and entraining signal we utilized. In fact, when we entrained with a heme shock we observed a very different response in terms of intracellular heme concentrations. The significance of the heme oscillations is unclear, but there are several possibilities that can be suggested including that the heme oscillations are required for mammalian clock function. However, this appears unlikely since we observed Per2 oscillations prior to initiation of intracellular heme oscillations (Fig. 3) suggesting that the oscillations in intracellular heme are not required for oscillations of at least some components of the mammalian clock. More likely is the possibility that intracellular heme levels offer a point of access for modification of the characteristics of the circadian rhythm. There are a number of examples of regulatory inputs that modify Alas1 expression including bile acids [24], nutritional status via PGC-1α [25], and xenobiotics [26,27]. Additionally, regulation of heme oxygenase, which controls heme degradation, is another mechanism that may be used to modify the mammalian clock via modulation of intracellular heme levels.

Fig. 4.

Model comparing the oscillations in Rev-erbα and intracellular [Heme] suggesting the possibility that there may be variability in the level of receptor occupancy and thus repressor activity while REV-ERBα is elevated during the circadian rhythm.