D. Rozman, Institute of Biochemistry, Center for Functional Genomics and Biochips, Faculty of Medicine, University of Ljubljana, Zaloska 4, SI-1000 Ljubljana, Slovenia Fax: +386 1 543 7588 Tel: +386 1 543 7591 E-mail: firstname.lastname@example.org
The cytochrome P450 (CYP) genes Cyp51, Cyp11a1, Cyp17a1, Cyb11b1, Cyp11b2 and Cyp21a1 are involved in the adrenal production of corticosteroids, whose circulating levels are circadian. cAMP signaling plays an important role in adrenal steroidogenesis. By using cAMP responsive element modulator (Crem) knockout mice, we show that CREM isoforms contribute to circadian expression of steroidogenic CYPs in the mouse adrenal gland. Most striking was the CREM-dependent hypomethylation of the Cyp17a1 promoter at zeitgeber time 12, which resulted in higher Cyp17a1 mRNA and protein expression in the knockout adrenal glands. The data indicate that products of the Crem gene control the epigenetic repression of Cyp17 in mouse adrenal glands.
Cytochromes P450 (CYPs) constitute a large superfamily of enzymes that perform biotransformations of different endogenous and exogenous substrates . Their tissue-specific and cell-specific expression is especially prominent in the adrenal gland . Here, Cyp51, Cyp11a1, Cyp17a1, Cyp11b1, Cyp11b2 and Cyp21a1 participate in the production of glucocorticoids (GCs), mineralocorticoids, and adrenal androgens, three classes of steroid hormones that are important in the regulation of metabolism, water and salt balance, and reproduction . Many CYPs, such as the CYP17 family, also exhibit species-specific expression. For example, the human adrenal cortex expresses CYP17A1 and can therefore produce cortisol, whereas rodents that do not express this gene in the adrenal glands produce corticosterone (Fig. 1) .
It is now widely accepted that the adrenal gland contains an intact circadian clock. Together with photic and nonphotic stimuli from the suprachiasmatic nucleus, this clock regulates the expression of about 5% of the adrenal genome. The circadian expression of genes involved in hormone synthesis results in the rhythmic expression of GCs observed in plasma of rodents and humans [5,6]. As GCs are known to regulate a wide array of different physiological processes, it has been suggested that they could play a central role as signaling molecules for the resetting of peripheral clocks [7,8]. If this is true, the deregulation of their synthesis could, in principle, lead to the disruption of peripheral circadian clocks and further contribute to the development of metabolic disorders, such as the metabolic syndrome [9,10].
The regulation of steroidogenic CYPs in the adrenal is under the tight control of adrenocorticotrophic hormone (ACTH). Binding of this hormone to the ACTH receptor elevates intracellular cAMP levels, leading to the activation of steroidogenic enzymes . The upregulation of cAMP also leads to activation of protein kinase A and the subsequent activation of the cAMP response element (CRE)-binding protein (CREB)/cAMP-responsive element modulator (CREM)/ATF-1 family of transcription factors. Among them, CREM has been shown to be involved in circadian regulation of melatonin production  and cholesterol synthesis . The role of CREM isoforms in circadian regulation of metabolic processes is well documented for melatonin synthesis in the pineal gland . Using a Crem knockout mouse model, we show for the first time that Crem contributes to the circadian expression of steroidogenic Cyp genes in the mouse adrenal gland and is involved in epigenetic repression of Cyp17.
Expression of steroidogenic Cyp genes in Crem−/− mice
The role of CREM in circadian expression of adrenal Cyp genes was evaluated by comparing the expression profiles of individual genes between wild-type (WT) and knockout (KO) mice (Fig. 2). A minimum of three biological replicates per time point were used. Although a clear and robust circadian pattern was seen for the core clock genes (data not shown), this pattern was less pronounced for adrenal Cyp genes (Fig. 2). The two-way analysis of variance showed a difference in expression between KO and WT animals for Cyp17a1, Cyp51, and Cyp21a1. The largest difference in amplitude was seen for Cyp17a1 at zeitgeber time (ZT) 12.
Methylation pattern of Cyp17 in Crem KO adrenals
DNA methylation is an important epigenetic factor that effectively prevents transcription of genes that are methylated . According to the literature, Cyp17 is methylated in rat adrenal glands and is therefore not expressed . As we observed higher-level expression of Cyp17 in mouse Crem KO adrenal glands, we examined its methylation pattern. Two candidate CpG islands reside in the mouse Cyp17 promoter. Their methylation status was evaluated in genomic DNA of Crem KO and WT mice by the use of methylation-sensitive BsrFI endonuclease in at least three replicates (Fig. 3A). If DNA is methylated, DNA is not cleaved, and the PCR product is observed on the gel. Figure 3B shows representative gels. The average of the three replicates is shown in Fig. 3C. We found that DNA was less methylated in Cyp17 CpG island 1 of the Crem KO mice, whereas no difference was observed in CpG island 2. Methylation in WT adrenals was not significantly different in either of the CpG islands. The upregulated expression of Cyp17a1 at ZT 12 was additionally confirmed by higher amounts of CYP17A1 as detected by western blotting (Fig. 3D).
CREM regulation of Cyp17a1
The Cyp17 promoter contains CREs (Fig. 4A) and is transcriptionally regulated through cAMP signaling. Here, we found that CREM transcription factors were also able to influence Cyp17 expression (Fig. 4B). The CREMτ activator, which is normally present only in the testis, includes the DNA-binding domains that appear in all Crem transcripts and bind to CREs. The data (Fig. 4B) suggest that the family of CREM proteins can activate Cyp17 transcription. One of the CREM isoforms is inducible cAMP early repressor (ICER), arising from one of the shortest Crem transcripts. It is considered to be the only cAMP-inducible transcription factor, and acts as a repressor. The maximum expression of Icer mRNA in WT animals was seen at ZT 12 (Fig. 4C). This was the same time when the difference in Cyp17a1 expression was observed between Crem KO and WT animals (Fig. 2). However, overexpression of ICER in Y-1 cells did not diminish the activity of the Cyp1a17 promoter-reporter (data not shown). Additional studies are required to clarify this issue and uncover the molecular mechanisms of the CREM-dependent regulation of mouse adrenal Cyp17a1.
Plasma corticosterone analysis
Because of elevated expression of Cyp17a1 in Crem KO mice at ZT 12, a difference in plasma corticosteroid concentration was expected. The plasma corticosteroids were initially analyzed by ELISA, with antibodies that are most specific to corticosterone, and later also by LC with tandem MS. Figure 5A shows a clear circadian rhythm of corticosterone in both WT and KO animals, but also suggests a trend for KO animals to have higher corticosterone levels. The difference was statistically significant at ZT 0, ZT 8 and ZT 16 as measured by ELISA, where five samples per group were investigated. The LC-MS/MS analysis (three samples per group) shows a nonsignificant trend for KO animals to have higher corticosterone levels at ZT 0, ZT 4, ZT 8, and ZT 24 (Fig. 5B). The differences between ELISA and LC-MS/MS can be explained by antibody crossreactivity with other plasma sterols in ELISA assays. It is thus possible that, in addition to corticosterone, other steroids are overexpressed in Crem KO adrenal glands. To date, we are unable to confirm or to exclude the presence of cortisol in plasma of Crem KO mice.
The adrenal gland seems to be the most important peripheral circadian oscillator in rodents. It mediates circadian signals from the suprachiasmatic nucleus to peripheral organs through production of GCs [7,8,16]. If the adrenal circadian rhythm is disrupted, the synchronization of peripheral organs is lost, and this can lead to the development of metabolic syndrome . In the mouse, corticosterone is the major adrenal corticosteroid produced in a circadian manner in response to ACTH oscillation . Cyp51, Cyp11a1, Cyp17a1, Cyp11b1, Cyp11b2 and Cyp21a1 are involved in the production of corticosteroids, where cAMP signaling is frequently involved in Cyp gene regulation . Therefore, we examined the circadian expression of Cyp genes involved in steroid metabolism in adrenal glands of Crem KO and WT mice, in addition to Cyp39 and Cyp7b1 from the oxysterol pathways. Because light can influence the expression of genes through cAMP signaling , mice were killed in dark-room conditions, as explained in Experimental procedures. In this way, the direct influence of light is excluded, and only changes resulting from regulation by the core clock genes are observed. Three of the measured Cyp genes (Cyp17a1, Cyp51, and Cyp21a1) show low-amplitude circadian expression that differs between Crem KO and WT mice. The most evident change is the increase in Cyp17a1 expression in Crem KO adrenal glands at ZT 12 (Fig. 2). The remaining Cyp genes (Cyp11a1, Cyp11b1, Cyp11b2, Cyp39a1, and Cyp7b1) fail to show a circadian expression profile in the WT mouse adrenal gland.
The reasons for the circadian upregulation of Cyp17a1 in Crem KO mice might be transcriptional, post-transcriptional, or epigenetic. It is known that the rat Cyp17 promoter is hypermethylated in adrenal glands as compared with the testis . This is the reason why rodents synthesize corticosterone as the main GC, whereas the level of cortisol, the major human GC, is below the level of detection . Methylation analysis indicates that the Cyp17a1 promoter is also hypermethylated in the mouse WT adrenal gland. Hypomethylation of the promoter CpG island 1 in Crem KO mice is in line with the observed higher Cyp17a1 mRNA and CYP17A1 expression in KO mice than in WT mice. We used the WT testis as a positive control, as CYP17A1 is expressed in Leydig cells, being involved in testosterone production .
The majority of Cyp genes from the adrenal gland are subject to cAMP regulation . Human CYP17 is activated by cAMP through four CREs in its proximal promoter [22,23]. CREM is a common name for a set of transcription factors that act as activators or repressors and can contain one or both DNA-binding domains encoded by the Crem gene [24,25]. CREMτ is normally not present in adrenal glands, but is known to be one of the strongest Crem activators . To determine in vitro whether CREM isoforms can trans-activate mouse Cyp17, we applied CREMτ, as it contains both DNA-binding domains. The regulation of mouse Cyp17a1 by CREM was proven to be activation, the intensity of the response being in the same range as for CREB. As Crem is alternatively spliced, it is difficult to design quantitative PCR (qPCR) assays to measure individual transcripts. However, we managed to measure the shortest transcript, Icer, the only cAMP-inducible form in the Crem family. ICER has been described as an early response gene and as a repressor, responding to cAMP stimuli through binding to CRE sequences and replacing other CRE_binding activators, such as CREB . Icer is strongly expressed in various regions of the brain, whereas its expression and regulatory roles in peripheral organs are not well understood . Our data show that the expression of Icer mRNA is circadian in adrenal glands and peaks at ZT 12. In addition to demethylation of the Cyp17a1 promoter, this would aid in explaining the upregulation of Cyp17 in Crem KOs at ZT 12. However, overexpression of ICER in transfection studies failed to repress Cyp17a1. We have to consider that Icer is prone to autoregulation , and that a small amount of basal transcription of Icer occurs in the Y-1 cell line that was used for transfection studies. Other possibilities include indirect actions of ICER by modulating factors that contribute to transcriptional, post-transcriptional or epigenetic regulation of Cyp17.
As discussed earlier, corticosterone is circadian in mouse plasma , and the maximum at ZT 12  was confirmed in our study in WT and Crem KO mice. The two analytical techniques applied (ELISA and LC-MS/MS) showed a trend of elevated corticosterone in KO mice at certain time points; however, a statistical threshold was reached only at ZT 0, ZT 8 and ZT 16 when ELISA was used. As the antibodies applied in ELISA exhibit crossreactivity with other plasma steroids, we cannot conclude that corticosterone is the only elevated steroid. According to LC-MS/MS, it seems that the circadian concentration of plasma corticosterone is not significantly influenced by the absence of Crem. However, ELISA indicates that the circadian expression of corticosterone and/or the crossreacting steroids might be elevated by the absence of Crem, especially at certain time points. Detailed LC-MS/MS steroid profiling will be used in the future to address this issue.
The increased corticosteroid concentration in Crem KO mice cannot be explained solely by changes in the transcription of steroidogenic Cyp genes in the adrenal glands. It is known that cAMP regulation in humans leads to phosphorylation of CYP17 by protein kinase A . This rapid modulation of activity could either directly affect CYP17 enzymatic activity [31,32] or could change the activity of the transcription factors responsible for its regulation . Another possibility is epigenetic regulation, as shown in our study, where the mouse Cyp17 promoter was less methylated in the absence of Crem.
In conclusion, this is the first report revealing the in vivo role of CREM transcription factors in the circadian and epigenetic regulation of mouse adrenal steroidogenic Cyp genes, in particular of Cyp17a1.
Circadian sample collection and RNA isolation
Fifty-four WT and 45 Crem KO mice were used. Animals had free access to food and water, and were maintained under a 12 : 12-h light cycle (light on at 7:00 a.m.; light off at 7:00 p.m.). The experiment was approved by the Veterinary Administration of the Republic of Slovenia (license number 34401-9/2008/4) and was conducted in accordance with the European Convention for the protection of vertebrate animals used for experimental and other scientific purposes (ETS 123), as well as in accordance with National Institutes of Health guidelines for work with laboratory animals. The process of sample preparation and total RNA isolation has been described previously . In brief, mice were killed under dim red light every 4 h during a 24-h period starting at 7 a.m. (ZT 0). The liver and adrenal glands were excised, snap frozen in liquid nitrogen, and stored at −80 °C.
Primer design and qPCR analysis
Intron-spanning primers were designed for all Cyp genes (Table 1A). Primer specificity and amplification efficiency were also validated empirically by melting curve and standard curve analysis of a six-fold dilution series. Primers and probe for Icer were designed by PrimerDesign (Southampton, UK). Normalization was performed according to , with Eq-PCR Wizard. Real-time qPCR was performed in a 384-well format on a LightCycler 480 (Roche Applied Science, Penzberg, Germany), with LightCycler 480 SYBR Green I Master Mix for Cyp genes and LightCycler 480 Probes Master Mix for the Icer gene (Roche Applied Science, Penzberg, Germany). The PCR reaction consisted of 2.5 μL of Master Mix, 1.15 μL of RNase-free water, 0.6 μL of 300 nm primer mix and 0.75 μL of cDNA in a total volume of 5 μL. Three technical replicates were performed for each sample. Cycling conditions were as follows: 10 min at 95 °C, followed by 40 rounds of 10 s at 95 °C, 20 s at 60 °C, and 20 s at 72 °C. Melting curve analysis for determination of the dissociation of PCR products was performed from 65 to 95 °C. Two-way analysis of variance was used for determination of statistically significant differences between KO and WT mice.
Table 1. Primer information for (A) qPCR and (B) methylation analysis. fw, forward; rv, reverse.
a Number of duplicated amplicones per cycle. Maximum efficiency is 2.
Mouse Cyp17a1 (accession number AY594330) was analyzed for the presence of CpG islands with the cpg island searcher program (http://cpgislands.usc.edu/). The CpG island was defined as a DNA sequence of 200 bp with a calculated percentage of CpGs higher than 50% and a calculated versus expected distribution of more than 0.6.
DNA was isolated from the same samples as RNA with Tri Reagent (Sigma, St. Louis, MO, USA), according to the manufacturer’s instructions. Five Crem KO and three WT samples were used. One microgram of DNA was incubated with the methylation-sensitive endonuclease BsrFI (Cfr10I; Fermentas, St. Leon-Rot, Germany) for 48 h. Fresh enzyme was added to the reaction every 24 h. DNA was purified with the QIAquick PCR Purification kit (Qiagene, Hilden, Germany) and used for PCR with the CpG-specific primers. β-Actin primers were used as an internal control (Table 1B). Cycling conditions were as follows: 5 min at 95 °C, followed by 35 rounds of 1 min at 95 °C, 30 s at 55 °C and 30 s at 72 °C, and a final extension for 5 min at 72 °C. PCR products were visualized on a 1.5% agarose gel with ethidium bromide staining and detected on an LAS4000 (FujiFilm, Tokyo, Japan). Bands on a gel were quantified with multi gauge software (FujiFilm, Tokyo, Japan). T-tests were used for calculation of differences between untreated and samples treated with the methylation-sensitive endonuclease.
Protein isolation and western blotting
Proteins were isolated from the same samples as RNA and DNA with Tri Reagent (Sigma, St. Louis, MO, USA), according to the manufacturer’s instructions. Six Crem KO and six WT adrenal samples were used. As a control, proteins were isolated from a single WT testis with the same protocol as used for the adrenal samples. Anti-CYP17 IgG (Abcam, Cambridge, UK) and secondary antibodies [donkey anti-(goat HRP-conjugated IgG)] (Santa Cruz Biotechnology, Heidelberg, Germany) were applied. For a loading control, monoclonal anti-β-actin IgG1 (Sigma, St. Louis, MO, USA) with secondary anti-mouse HRP-conjugated IgG (Amersham, Freiburg, Germany) were used. Detection of CYP17A1 was performed with West Femto substrate (Pierce, Rockford, IL, USA). For detection of actin, we used ECL substrate (Pierce, Rockford, IL, USA). Detection was performed on an LAS4000 (FujiFilm, Tokyo, Japan). Bands on a gel were quantified with multi gauge software (FujiFilm, Tokyo, Japan). T-tests were used to determine statistically significant differences between Crem KO and WT samples.
Cell cultures and transfection
Mouse adrenal cortex cell line Y-1 (ATCC no. CCL-79) was grown in DMEM (Sigma D7777 Sigmaaldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (Sigma F9665 Sigmaaldrich, St. Louis, MO, USA). One day before the transfection, cells were subcloned onto 96-well plates at a density of 0.5 × 106 cells·mL−1. Lipofectamine 2000 Reagent (Invitrogen, Carlsbad, CA, USA) was used for transient transfection, together with a luciferase reporter construct containing −1041 bp of the mouse Cyp17 promoter (gift from C. E. Flueck  Universität Bern, Departement Klinische Forschung KiKl G3 812, Inselspital, CH-3010 Bern, Schweiz). Cells were cotransfected with expression vectors: pRSV Creb (gift from M. R. Waterman , Vanderbilt University Medical Center, 607 Light Hall, Nashville, TN 37232-0146, USA) and pCMV Crem Tau (gift from P. Sassone Corsi , Department of Pharmacology, 2115 Gillespie Neuroscience, University of California, Irvine, California 92697-4625, USA). For each transfection combination, 12 replicates were performed. After 48 h, the medium was removed, and cells were lysed with Promegas lysis buffer (Promega, Fitchburg, WI, USA). Luminiscence was measured with a ONE-Glo kit (Promega, Fitchburg, WI, USA) on a VictorX5 plate reader (PerkinElmer, Waltham, MA, USA). For evaluation of transfection efficiency, cells were cotransfected with pSV β-galactosidase plasmid as previously described , and absorbance was measured with Epoch (BioTek, Winooski, VT, USA).
Plasma corticosterone measurement
Blood was collected in EDTA-coated vials and centrifuged at 4000 g for 20 min at 4 °C. Plasma was removed and kept at −80 °C until the corticosterone concentration was determined with the Corticosterone EIA kit (Enzo Life Science, Farmingdale, NY, USA), following the manufacturer’s instructions. Five samples per time point were used for each group (WT and KO). Plasma corticosterone was also measured by LC-MS/MS, as previously described . We used three mouse samples per time point for each mouse group. Two-way analysis of variance was used to compare differences between KO and WT mice.
This work was supported by the Slovenian Research Agency program, grant P1-0104. R. Kosir and U. Prosenc are funded as young researchers by the Slovenian Research Agency. We would like to thank our collaborators for kind gifts of plasmids: C. E. Flueck for pGL3 with the mouse Cyp17 promoter, M. R. Waterman for the expression vector pRSV Creb, and P. Sassone Corsi for the expression vector pCMV Crem Tau.