| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Zoology, School of Biological Sciences, Aberdeen University, Aberdeen, UK;
Cardiff School of Biosciences, Cardiff, UK;
University of Surrey, Guildford, UK; and
Centre for Reproductive Biology, University of Edinburgh, Queen’s Medical Research Institute, Edinburgh, UK
1Correspondence: Aberdeen University, School of Biological Sciences, Zoology, Tillydrone Avenue, Aberdeen, AB242TZ, Scotland, UK. E-mail: d.hazlerigg{at}abdn.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: pars tuberalis • clock gene • cryptochrome • photoperiod • circadian
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and plays an essential role in the control of seasonal physiology and behavior in mammals, acting as a humoral encoding of photoperiodic information (2)
. Melatonin acts on sites within the hypothalamus and pituitary to exert its seasonal effects, and the pars tuberalis (PT) of the hypophysis consistently expresses the highest concentration of mt1 receptors of all mammalian tissues, implying that it is a principal site of melatonin action (3)
.
Efforts to understand the molecular mechanisms of melatonin action show that the clock genes per1, per2, cry1, rev-erb
, and bmal1 are all rhythmically expressed in the PT, and their expression is modulated by melatonin (4)
. Two clock genes, in particular, per1 and cry1, appear to be acutely regulated in opposite ways. The rise of per1 expression at dawn is dependent on the declining plasma melatonin concentration, while the cry1 peak at dusk depends on its increase (4
5
6
7)
. This has led to the concept of a "coincidence timer"-based model for melatonin action, in which photoperiodic effects on the relative phasing of peaks of PER and CRY protein expression determine the transcriptional profile and hence the output of melatonin responsive cells (8)
.
Melatonin injection at the end of the night, extending the endogenous melatonin profile inhibits per1 expression in the morning (5)
, indicating that the morning per1 peak arises through a derepression mechanism when endogenous melatonin levels fall at dawn. Melatonin acutely inhibits cAMP production through a Gi type G-protein coupled signaling pathway (9)
, while prolonged exposure to melatonin causes sensitization of adenylate cyclase to stimulation when melatonin is withdrawn (10)
. These two effects jointly account for per1 regulation in PT cells (11
, 12)
.
In contrast, there is no clear evidence to link cry1 induction to the cAMP signaling pathway, in PT cells or in any other context. Exogenous administration of melatonin in rat or sheep at times of low or absent endogenous melatonin leads to the acute induction of cry1 in the PT (4
, 7)
. This effect appears to be independent of the time after lights on at which melatonin is administered, implying that the cry1 response to melatonin is independent of any circadian-based gating mechanism. Moreover, the magnitude of peak cry1 expression following melatonin injection, or evening rise in endogenous levels, greatly exceeds that directly due to circadian oscillator mechanisms running in the PT in the absence of melatonin (4)
. Hence, in the PT, melatonin is the dominant influence on cry1 transcriptional control.
So far as we are aware, beyond the PT, there are no other examples in which cry1 is dominantly regulated by a receptor-mediated input signal. A recent microarray study identified only three additional genes whose expression was, like cry1, significantly up-regulated (but with a lower magnitude) in the early night and acutely induced by melatonin: NeuroD1 (neurogenic differentiation factor 1), Pbef//Nampt (nicotinamide phosphoribosyltransferase) and Hif1a (hypoxia-inducible factor-1alpha) (13)
. Despite this growing list of melatonin-induced genes, the mechanism behind such induction remains unknown.
Here, we report that in addition to cry1, the melatonin receptor mt1 is also acutely and phase independently controlled by melatonin in vivo. Interestingly mt1 expression regulation is opposite to that of cry1. This inverse regulation can be recapitulated in vitro, revealing differential cAMP-dependence in the responses of these two genes to melatonin. Further, we used a protein-DNA interaction array to screen for novel transcriptional pathways involved in the evening effects of melatonin. This and follow-up experiments implicate the immediate early gene egr1 in melatonin-dependent transcriptional control in the PT.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Time of day effects on gene expression in the PT were assessed in tissue collected previously (12)
. Briefly, female Soay sheep were acclimated to an 8-h photoperiod (LD 8:16) for 8 wk and then were switched to LD 16:8 for a further 6 wk, prior to sacrifice and tissue collection at the time points indicated.
The effects of constant light and melatonin injections on mt1 gene expression were assessed in tissue collected previously (4)
. Briefly, sheep were acclimated to a long photoperiod (LD 16:8) for 6 wk. For the last 24 h, the lights were left on to suppress endogenous melatonin secretion. At 4-h intervals across 24 h, subcutaneous melatonin implants (Regulin; CEVA Animal Health, Chesham, Bucks, UK) were inserted on the inside of the hind leg using a standard injector. For sham treatment, the injector was left empty. Three hours after melatonin/sham treatment, sheep were killed by pentobarbital injection. Hypothalamic blocks with the anterior pituitary gland and PT on the ventral side were dissected within 10 min of death and snap frozen in –30°C isopentane.
For the Panomics array experiment, Soay sheep were acclimated to LD 16:8, and Regulin melatonin or sham treatments were administered 6 h after lights on. Animals were sacrificed 1.5 h postmelatonin/sham administration.
Culture of PT primary cells and explants
Reagents and materials were purchased from Sigma-Aldrich (Poole, UK) unless otherwise stated.
Cultures of ovine PT cells were prepared according to the method previously described (14)
. Cell count and viability were estimated according to the Trypan-blue exclusion method. For the culture of explants, sheep PT (all collected at a local abattoir during the morning hours, from 9 to 11 AM) were washed in ice-cold Gey’s Balanced Salt Solution (GBSS) with 0.6% glucose before being dissected into 2 torus-shaped (
2 mm thick) explants across the rostral/caudal axis. Explants were then transferred into separate T25 tissue culture flasks (Nunc, Fisher Scientific, Loughborough, UK) containing ice-cold GBSS 0.6% glucose. The GBSS was then decanted and replaced with 15 ml of ice-cold Supplemented Medium 199 (SM199, Invitrogen, Paisley, UK) with Earle’s salts, L-glutamine, 25 mM HEPES and L-amino acids (Invitrogen) supplemented with 0.6% glucose and antibiotic/ antimycotic (Invitrogen). Explants were precultivated for 1 h at 37°C, 5% CO2 on a rocking platform to allow gradual rewarming of the tissue. Medium was then decanted and replaced with fresh SM199 prewarmed to 37°C. After at least 8 h of culture, drug treatments were applied. This timing corresponds to late light phase or early dark phase relative to clock time experienced by donor sheep immediately prior to death. On the basis of preliminary experiments in which the expression of the forskolin-inducible clock gene, per1 (11)
was examined (data not shown), this 8-h preincubation also allowed a stable basal condition to be reached.
After culture, explants were carefully removed from the flasks, excess medium was removed, and the explants were embedded in TissueTek (RA Lamb, Eastbourne, UK) and quickly frozen. Explants were kept at –80°C until sectioning. Twenty-micrometer sections were cut using a Leica CM3050 S cryostat (Leica Microsystems, Wetzlar, Germany) and thaw-mounted onto poly-L-lysine-coated slides.
In situ hybridization
The expression of ovine cry1 and mt1 was analyzed by radioactive in situ hybridization, autoradiography, and densitometry using homologous mRNA probes, as described previously (8)
. GenBank accessions for the cDNA templates for ocry1 and omt1 are AY275673 and NM_001009725, respectively. For the explants, analysis of cry1 expression was limited to the area in which mt1 expression was observed. From mt1 in situ hybridization in the first adjacent sections, digital masks delimiting exactly the area of mt1 expression were created in ImageJ. This was pasted onto the corresponding cry1 autoradiograph as a transparency in ImageJ, and the threshold parameters were adjusted until the whole area defined by the mask was selected, from which optical density (OD) was quantified. Since the PT, as opposed to the median eminence or the hypothalamus, strongly expresses mt1, the mask so obtained defines the limits of the PT in each explant.
Sequencing and cloning of ovine cry1 5' regulatory region
A 1902-bp fragment of the 5' flanking region of the ocry1 gene (GenBank accession number EF651797), from –1801 to +101 relative to the predicted transcription start site was cloned and sequenced using the following primers: 5'-gcgggtaccgacgaccgaggatgagatg-3' and 5'-gcgaagctttcggcccgggtaagaga-3'. The truncated form of the promoter was cloned using the primers 5'-gcgggtacctcggagccgctgtagtaa-3' and 5'-gcgaagctttcggcccgggtaagaga-3' (–121 to +101). Amplicons were double-digested with KpnI and HindIII and cloned into pGL3-basic vector. Each PCR mix contained 10x PCR buffer, 5 µl, 0.2 mM of each dNTP, 2 mM MgCl2, 0.4 mM of each primer, 1 µl target DNA (50–100 ng), 0.2 µl of platinum Taq High Fidelity (Invitrogen), and DNase-free water to a final volume of 50 µl. Cycling parameters were initial denaturation at 95°C for 2 min followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 55°C for all primers for 30 s, elongation at 68°C for 1 min/kb of target DNA, followed by a final extension at 68°C for 10 min.
Extraction of nuclear proteins
Nuclear proteins were extracted using a Nuclear Extraction Kit (ActiveMotif, Rixensart, Belgium) following the manufacturer’s protocol except for the initial steps, modified for our purpose. Sheep PT tissue was dissected from the brain and directly immersed in ice-cold Gey’s BSS. A single PT was then transferred to a Kontes Dounce homogenizer containing 1 ml ice-cold 1x hypotonic buffer supplied with the kit [supplemented with 1 µl 1M dithiothreitol (DTT), 1 µl detergent, and 50 µl phosphatase inhibitor, all supplied with the kit], and homogenized by 20 strokes with a pestle (clearance type B). Subsequent steps were performed as described in the manufacturer’s protocol. Protein quantification was performed using the bicinchoninic acid (BCA) assay (Pierce, Fisher Scientific) with reduced volumes adapted for the NanoDrop Spectrophotometer ND-1000 (NanoDrop Technologies, Labtech, Ringmer, UK).
Electromobility shift assays
Double-stranded 30-mer oligonucleotides (MWG Biotech, London, UK), bearing a single nucleotide overhang at the 5' extremities, at a concentration of 5 µM were annealed by heating to 85°C for 5 min then left to cool. Oligonucleotides were labeled with 
32P-adenosine triphosphate (32P-ATP; GE Healthcare, Little Chalfont, UK). Three microliters of labeled oligonucleotides at 2 x 10E6 counts per minute (CPM)/µl was incubated with 20 µg of nuclear proteins, 12 µl of binding buffer (1 M MgCl2, 12.5 µl; 0.5 M EDTA, 2 µl; 1 M DTT, 1 µl; glycerol, 200 µl; 1 M Tris-HCl, pH 7.9, 50 µl; water, 735 µl), and 1 µl Poly(dI-dC). Poly(dI-dC) (GE Healthcare) at 0.5 µg/µl and water for a final volume of 24 µl. Binding reactions were incubated at room temperature for 20 min. Acrylamide gels (4%) were prepared and run on a Protean II Xi system with 1.5-mm-thick spacers and combs (Bio-Rad, Hemel Hempstead, UK). An initial prerun of the unloaded gels was performed at 105 V for 1 h in 0.25x tris-borate-EDTA. Then, samples were loaded and migrated at 240 V for 15 min, followed by a slower migration at 150 V for 3 h. Autoradiography films (MXB blue; Kodak, Rochester, NY, USA) were exposed to the gels for 3 to 24 h before development. Films were digitized using an Epson 1640Xl transmittance scanner (Epson UK Ltd., Hemel Hempstead, UK) and calibrated with Kodak optical density standards. Optical density (OD) measurements were performed using Image J freeware from the U.S. National Institutes of Health (NIH; Bethesda, MD, USA).
Protein/DNA interaction arrays
Protein/DNA Spin Combo arrays (Panomics, Vignate-Milano, Italy) were probed according to the manufacturer’s protocol. Enhanced chemiluminescence hyperfilms (GE Healthcare) were exposed to the membranes from 10 s to 5 min. Films were digitized as described above. Analysis of numerical data was performed according to the following procedure. For each array, individual values were normalized by dividing them by the sum of all the spot intensities on the array, giving the contribution of each spot to the total OD of the array. Normalized values were then averaged by treatment across the three replicates and plotted in a regression analysis using the values from one treatment against the values for the other treatment. Residual values for each data point were recorded. In addition, the fold-difference ratio for each value between the two treatments was calculated. Candidates were defined as those spots having a high residual value, over the 99% prediction band of the regression analysis, a high fold difference ratio (
2).
Cloning of the ovine egr1 expression vector
The ovine egr1 coding sequence was amplified by PCR using the primers 5'-ccgaagaattcatggcggcagccaaggc-3' and 5'-ccgttctcgagtggccatctcctcctcc-3', incorporating EcoRI and XhoI restriction overhangs, respectively. The double-digested PCR product was ligated into pCS2 (Invitrogen). The corresponding sequence has been deposited in GenBank under accession number EU552504.
Transient transfection and luciferase assays
COS-7 cells were grown in Dulbecco modified Eagle medium (DMEM) with L-glutamine, 4500 mg/L D-glucose, 25 mM HEPES, 1 mM sodium pyruvate, 1x antibiotic/antimycotic and 10% fetal bovine serum (all reagents from Invitrogen). Cells were transfected with the appropriate vectors using Genejuice reagent (Novagen, Merck, Nottingham, UK) according to the manufacturer’s protocol. For the luciferase assays, cells were washed 2 times with 1x phosphate buffer saline and lysed with 1x Reporter Lysis Buffer (Promega, Southampton, UK). Cell lysate (20 µl) was transferred to a clear-bottomed luminometer 96-well plate and 100 µl of luciferase reagent (Promega) was added. The plate was read by a Packard Lumicount luminometer (count duration of 2 s; Packard Instrument Co., Downers Grove, IL, USA). For the β-galactosidase assay, the remaining cell lysate was mixed with 900 µl of a buffer containing 60 mM Na2HPO4·7H2O, 40 mM NaH2PO4·H2O, 10 mM KCl, 1 mM MgSO4, 50 mM β-mercaptoethanol, and 1.3 mg/ml ortho-nitrophenyl-b-D-galactopyranoside. Absorbance at 420 nm was measured using a spectrophotometer (ND-1000; NanoDrop Technologies).
One-step quantitative real-time PCR
cDNA synthesis and real-time PCR were performed simultaneously using a LightCycler 1.5 and quantification software from Roche (Burgess Hill, UK) together with a Quantitect SYBR Green RT-PCR kit from Qiagen (Valencia, CA, USA), following the manufacturers’ protocols. Primers for the amplification of egr1 were 5'-ggcagaaggacaaaaaagc and 5'-agccgggagaggagtagga; for GAPDH, they were 5'-ggctgcccagaacatcatcc and 5'-ctccaggcggcaggtcag. GAPDH was found to be a suitable housekeeping gene with stable expression in our experiments. Common cycling parameters were used for all targets: reverse transcription at 50°C, 20 min; initial denaturation at 95°C, 15 min; followed by 40 cycles of denaturation at 95°C for 15 s, annealing at 55–60°C for 25 s, extension at 72°C for 20 s. Melting curve analysis used to check for specificity of amplification reactions. Real-time PCR analysis and crossing-point calculations were estimated using the second derivative maximum method.
Statistical analysis
Data are presented as means ± SE of three replicate experiments where appropriate. All statistical analyses were performed using Prism4 (GraphPad Software, San Diego, CA, USA). Data were analyzed by one- or two-way ANOVA, followed by Bonferroni post hoc testing as appropriate. Statistical significance was defined as P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). This pattern is inverse to that previously described for cry1.
|
We next analyzed mt1 expression in PT tissue from a previous study (4)
, in which long-day acclimated Soay sheep were exposed to constant light for up to 24 h, and given either sham (saline) or melatonin injections 3-h before each tissue sampling point (Fig. 1B
). In sham-injected control animals, mt1 and cry1 expression profiles exhibited low-amplitude oscillations persisting in the absence of melatonin and running in antiphase to each other (P<0.01 for effect of time for both genes, one-way ANOVA). In melatonin-injected animals, uniformly low levels of mt1 and high levels of cry1 expression were seen at all sampling points, akin to those seen shortly following lights off in animals held on a light-dark cycle (Fig. 1A
). Hence, mt1 appears to be an acutely melatonin-sensitive gene in the PT, with inverse control to that seen for cry1.
Cry1 regulation by melatonin in PT explants
Previous studies in ovine PT cells suggest that regulation of mt1 is cAMP dependent (15)
. To investigate the extent to which this also applies for the acute "evening type" effect of melatonin on cry1 expression, we cultured PT explants and treated them with the adenylate cyclase activator forskolin and/or melatonin (Fig. 2A
). Forskolin and/or melatonin treatment was administered at least 8 h after the start of the explant culture. Baseline cry1 expression was unaffected by forskolin treatment (5 µM). Melatonin (10 nM) caused a significant induction of cry1, and this effect was not significantly reduced by cotreatment with forskolin. In the same experiment, forskolin stimulated expression of mt1 in PT explants, and this effect was completely suppressed by cotreatment with melatonin (Fig. 2B
). Hence, mt1 and cry1 expression in the PT is differentially sensitive to cAMP.
|
Previous studies of morning per1 induction following melatonin withdrawal highlight a priming effect of melatonin exposure in the preceding hours (11)
. We wondered whether a similar, priming effect on evening cry1 might be exerted by prior ("morning") activation of the cAMP pathway. PT explants were pretreated with forskolin or vehicle control for 2 h, then washed and allowed to rest for 6 h prior to treatment with melatonin or vehicle (Fig. 2C, D
); data are expressed relative to control RNA expression in vehicle-pretreated explants. Forskolin pretreatment did not significantly increase cry1-responsiveness to melatonin, although there was a tendency toward this effect (Fig. 2C
; P=0.07 for treatmentxpretreatment interaction in 2-way ANOVA). Pretreatment did not have any significant effect on the inhibition of MT1 expression by melatonin (Fig. 2D
). Representative autoradiographic images of high and low Cry1 and MT1 expression in explants are shown in Fig. 2E
.
Melatonin-dependent changes in PT nuclear protein binding/activity
We used Panomics protein/DNA arrays to explore changes in DNA binding of nuclear proteins extracted from the PT of Soay sheep 1.5 h following melatonin or saline injections. A regression analysis of spot intensities in saline vs. melatonin-injected animals revealed several response elements showing significantly altered binding (Fig. 3A
). After filtering to remove spots in the arrays showing low intensities and high variability between replicates, maximally increased binding intensity in animals injected with melatonin was observed for the Panomics element EGR1 (a consensus element for the immediate early gene transcription factor egr1). Maximally reduced intensity following melatonin injection was seen for transcription factor building to immunoglobulin heavy chain enhancer 3, long isoform (TFE3L), which contains a tandem repeat of E-boxes (CAnnTG), and is known to bind to microopthalmia transcription factor 1 (MITF1) (16)
. Other outliers showing higher binding activity in melatonin-administered animals were AP3 (activator protein complex 3), E47 (immunoglobulin enhancer binding protein 47, a bHLH protein) and ISRE (interferon alpha-stimulated response element). Outlier candidates associated with lower binding activity in nuclear extracts from melatonin-treated animals were CREB-BP1 (bearing a response element for CREB-binding protein 1), HSE (oligonucleotide bearing a heat shock element), PAX8 (paired box 8 factor) and β-RE (bearing a response element for retinoic acid receptors). No obvious similarities in the sequence of the binding oligos in both groups were identified. These results are summarized in Table 1
.
|
|
Egr1 has been previously reported as being important for the regulation of hypophyseal hormones and the developmental loss of mt1 in neonatal pars distalis gonadotrophs (17
18
19)
, and we therefore investigated the significance of this result further.
We cloned 1.8 kb of the ovine cry1 promoter and used MatInspector to compare its organization with that of the ovine mt1 promoter, which we had cloned previously (19)
(Fig. 3B
). This revealed the presence of a putative egr1-response element (EGR1-RE) immediately upstream of the predicted transcription start site (TSS), as well as an E-box a further 60 bp upstream. Investigation in other mammals indicates that this E-Box is well conserved (not shown). In common with the ovine mt1 promoter, putative elements for the pituitary transcription factor pitx-1 were also observed, the most proximal of these lying 106 bp upstream of the cry1 TSS. pitx-1 is known to interact with egr1 to control the expression of mt1 and of the beta subunit of luteinizing hormone (βLH) (17
18
19)
.
To determine whether the observed response of the Panomics egr1 site to melatonin reflected effects at the putative cry1 EGR1-RE, we performed electromobility shift assays (EMSAs) using extracts from individual animals injected with either melatonin (n=3) or saline (n=3; Fig. 3C
, with M and C indicating extracts from melatonin or control). This revealed a doublet of bands showing low mobility and significantly increased intensity in melatonin-injected animals (t test, P<0.001; OD 0.49±0.02 and 0.77±0.004 for control and melatonin-administered animal, respectively).
A functional EGR1-RE in the cry1 proximal promoter
For EGR1-REs to account for inverse regulation of mt1 and cry1, we would predict opposed effects of egr1 at EGR-REs present in their promoter regions. To test this, we made cry1-luciferase reporter constructs for cotransfection experiments in COS7 cells (Fig. 4
). We observed, however, a similar response to that previously reported for the mt1 promoter: pitx-1 dramatically increased expression of a –1.8 kb cry1 construct, and this effect was completely reversed by cotransfection of egr1 (Fig. 4A
). Egr1 alone was without effect.
|
In a further experiment using a truncated promoter construct containing the proximal 121 bp upstream of the ocry1 TSS, which includes both the proximal PITX-1 element and the EGR1-RE, a similar pattern of response was also observed (Fig. 4B
). In addition, we explored possible interactions between the PITX-1 and EGR1 elements and the E-box present in this truncated reporter. Cotransfection of clock and bmal1 increased luciferase expression to a similar extent to that seen for pitx-1, and these two effects were additive. EGR1 was able to suppress, but not completely block, the combined inductive effects of PITX-1 and CLOCK/BMAL1 on this truncated construct. In addition, although PITX-1 induced the expression of both the full length and truncated reporters, the magnitude of this induction was lower for the truncated one, consistent with the loss of an additional putative PITX-1 element present at position –183 in the ovine cry1 promoter.
Circadian and melatonin-dependent regulation of egr1 expression in ovine PT
In PT tissue from Soay sheep acclimated to short photoperiod (LD 8:16) for 6 wk, egr1 levels changed in an approximately sinusoidal manner throughout the 24-h cycle, with minimal and maximal levels occurring at the beginning and end of the dark phase, respectively (Fig. 5A, B
; P<0.01 for effect of time, one-way ANOVA).
|
To test whether melatonin could directly affect egr1 expression, we performed experiments in primary PT cell cultures, using the same design as that in which we had observed melatonin induction of cry1 (Fig. 5C
). This revealed a direct inhibitory effect of melatonin on egr1 expression, which was independent of whether cells had been pretreated with forskolin (P<0.01 for effect of treatment, two-way ANOVA). Treatment with forskolin alone for 2 h increased egr1 expression 2-fold compared to basal levels (results not shown; 1.0±0.09 cf. 2.6±0.28; P<0.01, t test).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Two previous observations were important in dictating our focus on egr1. The first was the demonstration of light-dependent induction of egr1 in the suprachiasmatic nucleus (SCN) and periventricular hypothalamus (21
, 22)
as part of the photoperiod transduction pathway. The second was the connection between GnRH-induced egr1 expression and anterior pituitary function. While the cellular consequences of light-dependent egr1 expression in the SCN remain unclear (23
, 24)
, GnRH-dependent egr1 expression in the pituitary has been linked to postnatal changes in βLH and mt1 melatonin receptor expression (17
18
19
, 25
, 26)
. Indeed, the up-regulation of βLH and suppression of mt1 in the pituitary pars distalis have been linked to the onset of pulsatile GnRH secretion at puberty, where GnRH drives a high level of expression of egr1 in gonadotropes (17
, 19)
. This contrasts with the situation in the pituitary cells of the PT in mammals, wherein mt1 expression persists throughout adulthood. The data presented here show that egr1 is itself a melatonin-regulated gene in this tissue.
The results of the Panomics screening, in silico analysis of the ovine cry1 promoter, and EMSA experiments point to a role of egr1 in signaling in the PT. Treatment with melatonin for 1.5 h appears to cause a rapid change in protein DNA interaction at an EGR1-RE in the proximal region of the ovine cry1 promoter, consistent with the speed of melatonin-induced gene expression (13)
. Given the literature linking egr1 to pituitary gene regulation (17
, 27
28
29)
and the observed doublet of retarded bands in the EMSA experiments, it is probable that melatonin modulates the interactive effects of egr1 and other PT-expressed transcription factors at this response element. However, other egr family members and the Wilms’ tumor suppressor protein, WT1, also bind to EGR1-REs (30)
, and thus, it remains to be established which proteins contribute to the melatonin-dependent retarded EMSA complex and thus function in the transduction relay.
We found a putative proximal EGR1-RE in both cry1 and mt1, genes showing acute evening regulation by melatonin but with an inverse response. Egr1 modulates transcription through G/C-rich elements typically of the form CGKRGGC, where K is G or T and R is A or G, leading to either transcriptional activation or repression, depending on promoter context (31)
. We wondered, therefore, whether the differential regulation of cry1 and mt1 by melatonin, might be accounted for by differential actions of egr1 on their promoter regions. Instead, we observed similar control, with egr1 acting to repress the expression of both genes in COS7 cells transfected with a luciferase reporter. This result suggests that in the PT, factors additional to egr1 lead to the differential regulation of cry1 and mt1. In this regard, it is interesting that the NGF1A-induced binding proteins, NAB1 and NAB2, have been demonstrated to act either as corepressors or coactivators of egr1, depending on promoter context (32)
.
A further possibility suggested by the present study is that differential effects of the cAMP pathway on mt1 and cry1 are key to their differential regulation in vivo. In the present study and in work by Barrett and colleagues (33)
, forskolin induced the expression of mt1, presumably through activation of the cAMP pathway. Since forskolin also induces the expression of egr1 in PT cells, and yet mt1 expression is not inhibited, we infer that the direct positive effects of cAMP on mt1 expression may dominate over the indirect repressive effects of cAMP-induced egr1 expression.
In contrast, we could observe no acute effect of forskolin on cry1 expression, suggesting that direct stimulatory effects of cAMP at the cry1 promoter are of negligible importance. This does not exclude possible indirect effects on cry1 expression through the cAMP pathway, although we were unable to show a significant priming effect of a pulse of forskolin treatment 6 h before melatonin treatment in the current study. Given that melatonin directly suppresses egr1 in vitro and that a nadir in the diurnal rhythm of PT egr1 expression in the PT is achieved following lights off, a cAMP-dependent effect of melatonin on cry1 through egr1 could be envisaged.
We used the pituitary transcription factor pitx-1 to induce either mt1 or cry1 expression to allow repressive effects of egr1 to be observed in COS7 cells. Pitx-1 was originally described as a pan-pituitary transcription factor that appears early in adenohypophyseal cell lineage differentiation (34)
. Highest levels of pitx-1 expression are seen in the rostral tip of the developing anterior pituitary, which corresponds to the PT in the adult (35)
. Moreover, an enriched level of expression of PT pitx-1 expression continues into adult life. This feature of the PT may be a crucial factor in both maintaining high mt1 expression and leading to the high sensitivity of cry1 expression to mt1 receptor-mediated input.
So far as we are aware, this input sensitivity, as opposed to circadian regulation of cry1, has only been documented in the PT. This suggests that pitx-1 or other pituitary cell-type-specific transcription factors maintain the promoter region of cry1 gene in a state of heightened sensitivity to transcription factors mediating receptor and second-messenger pathway inputs. In this regard, note that the proximal cry1 E-box, linked to circadian control of cry1 (36)
, lies between the proximal PITX-1 and EGR1 elements in the cry1 promoter. We speculate that this arrangement in the proximal promoter constitutes a cell context-dependent cry1 regulatory cassette. In the PT, dominant control of cry1 expression occurs through melatonin signal transduction pathways, with weak circadian effects only discernible when melatonin production is suppressed, for example, by exposure to LL. By contrast, in tissues such as the SCN, wherein no direct effects of melatonin or other hormonal or neurotransmitter signals on cry1 expression have been reported, E-box control dominates because of the absence of tissue-specific factors necessary to maintain the cry1 gene sensitive to receptor signaling pathways. The contrasting roles of these different influences in the PT and SCN are summarized in Fig. 6
. In the PT, the sensitivity of cry1 to inputs has been related to the process of photoperiodic time measurement; it remains to be seen whether tissue-specific promoter control leads to similar input sensitivity in cry1 expression patterns in other contexts.
|
Similar to the study of Dupré et al. (13)
, our Panomics study and subsequent data also reveal that bHLH (TFE3L), POU (PITX1), and paired box (PAX8) factors might be important for the response to melatonin. While bHLH factors can reasonably be linked to the effect of melatonin on E-box-driven clock gene expression (4
, 8)
, the significance of paired box factors, as well as of HSE and β-RE, remains to be confirmed. In addition, the decrease in binding intensity of the CREB-BP1 candidate once more confirms the inhibitory effect of melatonin on the cAMP/CREB pathway.
In conclusion, the present study presents data showing functional, proximal EGR1-REs are present in both the cry1 and mt1 genes and that these exert repressive effects on gene expression. Rhythmical and melatonin-dependent changes in PT egr1 gene expression suggest that evening gene regulation by this hormone involves egr1-dependent signaling. Further studies are required to detail the control of egr1 expression in the PT and to define the proteins with which egr1 interacts to control evening gene expression.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 10, 2008. Accepted for publication October 23, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
